OCA Oracle Database SQL Exam Guide (Exam 1Z0 071) (Oracle Press)

User Manual:

Open the PDF directly: View PDF PDF.
Page Count: 790 [warning: Documents this large are best viewed by clicking the View PDF Link!]

Copyright©2017byMcGraw-HillEducation.Allrightsreserved.Exceptas
permittedundertheUnitedStatesCopyrightActof1976,nopartofthis
publicationmaybereproducedordistributedinanyformorbyanymeans,or
storedinadatabaseorretrievalsystem,withoutthepriorwrittenpermissionof
thepublisher,withtheexceptionthattheprogramlistingsmaybeentered,
stored,andexecutedinacomputersystem,buttheymaynotbereproducedfor
publication.
ISBN:978-1-25-958461-9
MHID:1-25-958461-5.
ThematerialinthiseBookalsoappearsintheprintversionofthistitle:ISBN:
978-1-25-958549-4,MHID:1-25-958549-2.
eBookconversionbycodeMantra
Version1.0
Alltrademarksaretrademarksoftheirrespectiveowners.Ratherthanputa
trademarksymbolaftereveryoccurrenceofatrademarkedname,weusenames
inaneditorialfashiononly,andtothebenefitofthetrademarkowner,withno
intentionofinfringementofthetrademark.Wheresuchdesignationsappearin
thisbook,theyhavebeenprintedwithinitialcaps.
McGraw-HillEducationeBooksareavailableatspecialquantitydiscountstouse
aspremiumsandsalespromotionsorforuseincorporatetrainingprograms.To
contactarepresentative,pleasevisittheContactUspageat
www.mhprofessional.com.
OracleisaregisteredtrademarkofOracleCorporationand/oritsaffiliates.All
othertrademarksarethepropertyoftheirrespectiveowners,andMcGraw-Hill
Educationmakesnoclaimofownershipbythementionofproductsthatcontain
thesemarks.
ScreendisplaysofcopyrightedOraclesoftwareprogramshavebeenreproduced
hereinwiththepermissionofOracleCorporationand/oritsaffiliates.
OracleCorporationdoesnotmakeanyrepresentationsorwarrantiesastothe
accuracy,adequacy,orcompletenessofanyinformationcontainedinthisWork,
andisnotresponsibleforanyerrorsoromissions.
TERMSOFUSE
ThisisacopyrightedworkandMcGraw-HillEducationanditslicensorsreserve
allrightsinandtothework.Useofthisworkissubjecttotheseterms.Exceptas
permittedundertheCopyrightActof1976andtherighttostoreandretrieveone
copyofthework,youmaynotdecompile,disassemble,reverseengineer,
reproduce,modify,createderivativeworksbasedupon,transmit,distribute,
disseminate,sell,publishorsublicensetheworkoranypartofitwithout
McGraw-HillEducation’spriorconsent.Youmayusetheworkforyourown
noncommercialandpersonaluse;anyotheruseoftheworkisstrictlyprohibited.
Yourrighttousetheworkmaybeterminatedifyoufailtocomplywiththese
terms.
THEWORKISPROVIDED“ASIS.”McGRAW-HILLEDUCATIONAND
ITSLICENSORSMAKENOGUARANTEESORWARRANTIESASTO
THEACCURACY,ADEQUACYORCOMPLETENESSOFORRESULTS
TOBEOBTAINEDFROMUSINGTHEWORK,INCLUDINGANY
INFORMATIONTHATCANBEACCESSEDTHROUGHTHEWORKVIA
HYPERLINKOROTHERWISE,ANDEXPRESSLYDISCLAIMANY
WARRANTY,EXPRESSORIMPLIED,INCLUDINGBUTNOTLIMITED
TOIMPLIEDWARRANTIESOFMERCHANTABILITYORFITNESSFOR
APARTICULARPURPOSE.McGraw-HillEducationanditslicensorsdonot
warrantorguaranteethatthefunctionscontainedintheworkwillmeetyour
requirementsorthatitsoperationwillbeuninterruptedorerrorfree.Neither
McGraw-HillEducationnoritslicensorsshallbeliabletoyouoranyoneelsefor
anyinaccuracy,errororomission,regardlessofcause,intheworkorforany
damagesresultingtherefrom.McGraw-HillEducationhasnoresponsibilityfor
thecontentofanyinformationaccessedthroughthework.Underno
circumstancesshallMcGraw-HillEducationand/oritslicensorsbeliableforany
indirect,incidental,special,punitive,consequentialorsimilardamagesthat
resultfromtheuseoforinabilitytousethework,evenifanyofthemhasbeen
advisedofthepossibilityofsuchdamages.Thislimitationofliabilityshallapply
toanyclaimorcausewhatsoeverwhethersuchclaimorcausearisesincontract,
tortorotherwise.
ToJimBauchspies,whogavememyfirstOraclejobandwhohasalwaysbeen
likeasecondfathertome.
ABOUTTHEAUTHOR
SteveO’Hearnisatechnologyconsultantwithexpertiseinthedesign,
development,andadministrationofvariousdataanddata-drivensystemsfor
suchclientsastheU.S.DefenseDepartment,NASAHQ,theFAA,theWorld
Bank,andmanyothers.HefirstbecameanOracleCertifiedProfessional(OCP)
in2001andisacertifiedOracleDatabaseSQLExpert.Hehasadegreein
businessadministrationwithaspecializationininformationprocessingfrom
GeorgeWashingtonUniversity,andhispostgraduateworkincludesadvanced
statisticsandthecompletionoftheFutureofe-GovernmentExecutiveEducation
trainingatHarvardUniversity’sKennedySchoolofGovernmentin2003.Heis
amemberofMensa.
Mr.O’HearnauthoredOCEOracleDatabaseSQLCertifiedExpertExam
Guide(Exam1Z0-047),theofficialcertificationexamguideforOracle’sfirst
expert-levelcertificationexam,andthecriticallyacclaimedOCPDeveloper:
PL/SQLProgramUnitsExamGuide,bothfromOraclePress.Hecontributedto
OracleWebApplications101fromOraclePressandOracle8ServerUnleashed.
HehasbeenfeaturedonOracleTVandOracleOpenWorld/JavaOne.Hehas
beenpublishedinavarietyofpublications,includingtheHarvardBusiness
ReviewandtheRecordoftheNationalPressClub,whereheisanofficially
recognizedsubject-matterexpertonthetopicsofdatabaseandinformation
technology.HehasbeenvicepresidentandconferencecoordinatorfortheMid-
AtlanticAssociationofOracleProfessionals.HeleadstheofficialJavaUser
Group(JUG)fortheBaltimoreWashingtonMetropolitanArea,possesses
certificationsinJavaandothertechnologies,andlovesworkingwithhisown
RaspberryPiand3Dprintingdevices.
Mr.O’HearnprovidesOracleandJavatechnologytrainingandtutoring
online.HeblogsatSkere9.com.Heinvitesanyandallinquiriesat
soh@corbinian.com.
AbouttheContributor
BobBrylaisanOracle9i,10g,11g,and12cCertifiedProfessionalwithmore
than20yearsofexperienceindatabasedesign,databaseapplication
development,training,andOracledatabaseadministration.Heistheprimary
OracleDBAanddatabasescientistatEpicinVerona,Wisconsin.Heisa
technicaleditorforanumberofOraclePressbooks,includingseveral
certificationstudyguidesforOracleDatabase10g,11g,and12c.Youcanalso
findhimonLynda.comteachingOracleDatabasefundamentalsinbite-sized
videos.Hehasalsobeenknowntowatchscience-fictionmovies,tinkerwith
Androiddevices,andrefinishantiquefurnitureinhissparetime.
AbouttheTechnicalEditor
ToddMeisterhasbeenworkingintheITindustryformorethan20years.He’s
beenatechnicaleditoronmorethan75booksrangingintopicfromSQLServer
tothe.NETFramework.Besidestechnicalediting,heistheassistantvice
president/chiefenterprisearchitectatBallStateUniversityinMuncie,Indiana.
HelivesincentralIndianawithhiswife,Kimberly,andtheirfivebright
children.
CONTENTSATAGLANCE
1OracleandStructuredQueryLanguage(SQL)
2UsingDDLStatementstoCreateandManageTables
3ManipulatingData
4RestrictingandSortingData
5UsingSingle-RowFunctionstoCustomizeOutput
6UsingConversionFunctionsandConditionalExpressions
7ReportingAggregatedDataUsingtheGroupFunctions
8DisplayingDatafromMultipleTables
9UsingSubqueriestoSolveQueries
10ManagingSchemaObjects
11UsingtheSetOperators
12ManagingObjectswithDataDictionaryViews
13ManipulatingLargeDataSets
14ControllingUserAccess
AAbouttheDownload
Glossary
Index
CONTENTS
Acknowledgments
Preface
Introduction
ExamReadinessChecklist
1OracleandStructuredQueryLanguage(SQL)
TheExam:AnOverview
WhattoDoandWhattoExpect
OracleSQLvs.ANSISQL
OracleSQLvs.OracleSQL*Plus
SQLFundamentalsIvs.SQLCertifiedAssociate
ConfirmAppropriateMaterialsforStudy
IdentifytheConnectionBetweenanERDandaRelationalDatabase
Entity-RelationshipDiagramsandDataModeling
RelationalDatabases
Many-to-ManyRelationships
DatabaseNormalization
ExplaintheRelationshipBetweenaDatabaseandSQL
DescribethePurposeofDDL
DescribethePurposeofDML
TransactionControlLanguage
BuildaSELECTStatementtoRetrieveDatafromanOracle
DatabaseTable
CertificationSummary
✓Two-MinuteDrill
Q&ASelfTest
SelfTestAnswers
2UsingDDLStatementstoCreateandManageTables
CategorizetheMainDatabaseObjects
WhatAreDatabaseObjects?
Schemas
CreateaSimpleTable
NamingaTableorOtherObject
TheSQLStatementCREATETABLE
ReviewtheTableStructure
ListtheDataTypesThatAreAvailableforColumns
Character
Numeric
Date
LargeObjects
ExplainHowConstraintsAreCreatedattheTimeofTableCreation
CreatingCONSTRAINTSintheCREATETABLEStatement
TheTypesofCONSTRAINTS
DropColumnsandSetColumnUNUSED
DroppingColumns
UNUSED
CreateandUseExternalTables
Benefits
CreatingExternalTables
CertificationSummary
✓Two-MinuteDrill
Q&ASelfTest
SelfTestAnswers
3ManipulatingData
TruncateData
RecursivelyTruncateChildTables
InsertRowsintoaTable
DefaultColumnList
EnumeratedColumnList
UpdateRowsinaTable
Expressions
Constraints
TheWHEREClause
DeleteRowsfromaTable
ControlTransactions
COMMIT
ROLLBACK
SAVEPOINT
ROLLBACKRevisited
CertificationSummary
✓Two-MinuteDrill
Q&ASelfTest
SelfTestAnswers
4RestrictingandSortingData
SorttheRowsThatAreRetrievedbyaQuery
ReferencebyName
Expressions
ReferencebyPosition
Combinations
ORDERBYandNULL
LimittheRowsThatAreRetrievedbyaQuery
TheWHEREClause
BooleanLogic
AdditionalWHEREClauseFeatures
AdditionalConcepts
UseAmpersandSubstitutiontoRestrictandSortOutputatRunTime
&
DEFINEandUNDEFINECommands
TheSETandSHOWCommands
ACCEPTandPROMPT
UsetheSQLRowLimitingClause
FETCH
WITHTIES
OFFSET
CertificationSummary
✓Two-MinuteDrill
Q&ASelfTest
SelfTestAnswers
5UsingSingle-RowFunctionstoCustomizeOutput
UseVariousTypesofFunctionsThatAreAvailableinSQL
CharacterFunctions
NumberFunctions
DateFunctions
OtherFunctions
UseCharacter,Number,Date,andAnalytical
(PERCENTILE_CONT,STDDEV,LAG,LEAD)FunctionsinSELECT
Statements
TheDUALTable
CharacterFunctions
NumericalFunctions
DateFunctions
AnalyticalFunctions
NestingFunctions
CertificationSummary
✓Two-MinuteDrill
Q&ASelfTest
SelfTestAnswers
6UsingConversionFunctionsandConditional
Expressions
DescribeVariousTypesofConversionFunctions
ExplicitandImplicitConversion
UsetheTO_CHAR,TO_NUMBER,andTO_DATEConversion
Functions
ConversionFunctions
AdditionalConversionFunctions
ApplyGeneralFunctionsandConditionalExpressionsinaSELECT
Statement
CASE
DECODE
NVL
NULLIF
CertificationSummary
✓Two-MinuteDrill
Q&ASelfTest
SelfTestAnswers
7ReportingAggregatedDataUsingtheGroupFunctions
DescribetheUseofGroupFunctions
COUNT
SUM
MIN,MAX
AVG
MEDIAN
RANK
DENSE_RANK
FIRST,LAST
Others
GroupDatabyUsingtheGROUPBYClause
MultipleColumns
ORDERBYRevisited
NestingFunctions
IncludeorExcludeGroupedRowsbyUsingtheHAVINGClause
CertificationSummary
✓Two-MinuteDrill
Q&ASelfTest
SelfTestAnswers
8DisplayingDatafromMultipleTables
DescribetheDifferentTypesofJoinsandTheirFeatures
TypesofJoins
UseSELECTStatementstoAccessDatafromMoreThanOneTable
UsingEquijoinsandNon-Equijoins
InnerJoins
UsingTableAliases
NaturalJoins
USING
MultitableJoins
Non-Equijoins
JoinaTabletoItselfbyUsingaSelf-Join
Self-ReferencingForeignKeys
Self-JoinSyntax
ViewDataThatGenerallyDoesNotMeetaJoinConditionbyUsing
OuterJoins
LEFTOUTERJOIN
RIGHTOUTERJOIN
FULLOUTERJOIN
FortheRecord:OracleOuterJoinSyntax:(+)
CertificationSummary
✓Two-MinuteDrill
Q&ASelfTest
SelfTestAnswers
9UsingSubqueriestoSolveQueries
DefineSubqueries
DescribetheTypesofProblemsSubqueriesCanSolve
DescribetheTypesofSubqueries
QueryDataUsingCorrelatedSubqueries
UpdateandDeleteRowsUsingCorrelatedSubqueries
UPDATEwithaCorrelatedSubquery
DELETEwithaCorrelatedSubquery
UsetheEXISTSandNOTEXISTSOperators
UsetheWITHClause
WriteSingle-RowandMultiple-RowSubqueries
Single-RowSubqueries
Multiple-RowSubqueries
CertificationSummary
✓Two-MinuteDrill
Q&ASelfTest
SelfTestAnswers
10ManagingSchemaObjects
DescribeHowSchemaObjectsWork
Tables
Constraints
Views
Indexes
Sequences
CreateSimpleandComplexViewswithVisible/InvisibleColumns
CreatingViews
UpdatableViews
InlineViews
ALTERVIEW
Visible/InvisibleColumns
Create,Maintain,andUseSequences
CreatingandDroppingSequences
UsingSequences
CreateandMaintainIndexesIncludingInvisibleIndexesand
MultipleIndexesontheSameColumns
TheOracleDatabaseOptimizer
ImplicitIndexCreation
SingleColumn
Composite
Unique
Dropping
VisibleandInvisibleIndexes
IndexAlternativesontheSameColumnSet
PerformFlashbackOperations
Overview
RecoverDroppedTables
RecoveringDataWithinExistingTablesoverTime
MarkingTime
CertificationSummary
✓Two-MinuteDrill
Q&ASelfTest
SelfTestAnswers
11UsingtheSetOperators
DescribeSetOperators
UseaSetOperatortoCombineMultipleQueriesintoaSingleQuery
UNION
UNIONALL
INTERSECT
MINUS
Combinations
ControltheOrderofRowsReturned
ORDERBY—ByPosition
ORDERBY—ByReference
CertificationSummary
✓Two-MinuteDrill
Q&ASelfTest
SelfTestAnswers
12ManagingObjectswithDataDictionaryViews
QueryVariousDataDictionaryViews
Structure
DynamicPerformanceViews
ReadingComments
AddingComments
DICTIONARY
IdentifyingaUser’sOwnedObjects
InspectingTablesandColumns
CompilingViews
CheckingPrivileges
InspectingConstraints
FindingColumns
CertificationSummary
✓Two-MinuteDrill
Q&ASelfTest
SelfTestAnswers
13ManipulatingLargeDataSets
DescribetheFeaturesofMultitableINSERTs
UsetheFollowingTypesofMultitableINSERTS:
UnconditionalandConditional
MergeRowsintoaTable
CertificationSummary
✓Two-MinuteDrill
Q&ASelfTest
SelfTestAnswers
14ControllingUserAccess
DifferentiateSystemPrivilegesfromObjectPrivileges
SystemPrivileges
Prerequisites
GRANTandREVOKE
ANY
ADMINOPTION
ALLPRIVILEGES
PUBLIC
GrantPrivilegesonTablesandonaUser
SchemaPrefixes
WITHGRANTOPTION
REVOKE
ALLPRIVILEGES
DependentPrivileges
ViewPrivilegesintheDataDictionary
GrantRoles
DistinguishBetweenPrivilegesandRoles
CertificationSummary
✓Two-MinuteDrill
Q&ASelfTest
SelfTestAnswers
AAbouttheDownload
SystemRequirements
InstallingandRunningTotalTester
AboutTotalTester
TechnicalSupport
Glossary
Index
F
ACKNOWLEDGMENTS
irst,Iwanttoacknowledgeyou,thereader.Wereitnotforyou,thisbook
wouldneverhavebeencreated.Becauseofyou,becauseofyourinterest,
becauseofyourprofessionalism,andbecauseofyourdemonstrablecommitment
toimprovingyourselfandthereforetheindustry,thisbookwascreated.Wereit
notforyou,therewouldnotbeaneliteteamofexpertshereatOraclePress
respondingtoyourinterestandyourdemandforqualitymaterial.So,thankyou.
Withoutyou,wewouldnotbedoingwhatwedo,norwouldthisbookexist.
Speakingofthisbookthatyouareholdinginyourhands(orperhapsreading
onyourscreen),itisatremendousworkandavaluablecontributiontothe
industry.Itrepresentsagreatdealofeffortbyanumberoftalented,persistent,
anddedicatedprofessionals.
Thisbookoriginallystartedasasecondeditiontomy2009SQLExpertbook
onthe1Z0-047examandwasnearingcompletionwhenitevolvedintothis
book,thefirstcertificationexamguideforthe1Z0-071exam.
First,aspecialandhugethank-youtoHilaryFlood.Thisbookwouldnot
havehappenedwithouther.Itwasherbold,openmindandcreativeforcethat
helpedoriginallysteerthisshipthroughmurkywaterstoitseventualgoal.For
herpersistence,patience,expertise,andinsight,Ithankher.
ClaireYeewastheeditorialcoordinator.Shewasconsistentlycheerfuland
encouragingandworkedtirelesslytoensureeachchapterorothersection
workeditswaythroughthemanyeditorialcycles,fromoriginalmaterialthrough
technicaledit,copyediting,andbeyondtofinalproduction.
DipikaRungtawastheprojectmanager.Asthevariouschaptersandother
sectionsexitedtheprocessheadedbyClaire,eachmovedtoDipikafor
additionalreview,work,andthemanydemandingdetailsnecessarytoprepare
thebookforproduction.Dipikahasbeenajoytoworkwithandisatremendous
assettoanybookproductioneffort.
JodyMcKenziewastheeditorialsupervisorandsteppedintoresolvesome
challengingissuesatkeytimes.LisaMcCoywastheproofreaderandcaught
someofthemostobscuretyposimaginable.ThecopyeditorwasKimWimpsett
andhelpedmakesuresomeofmycuriousexpressionsorwordchoiceswere
translatedintounderstandablenarratives.TheproductionsupervisorwasJim
Kussow,whohasmanagedtheproductionandprintingprocessesofthis
ponderoustome.
ThankyoutoVasundharaSawhneywhocontributedalotofworkinthe
earlieststagesofthisjourney.
TheacquisitionseditorwasLisaMcClain.Acquisitionseditorsleadthe
chargetoputbookprojectstogetherinthefirstplaceandstayinvolvedtoensure
thoseprojectsmarchforwardtotheirdestination.Theymaketoughcallsinthe
faceofuncertaintyandhavetodiscernapathforwardwithsometimeslimited
information.OraclePresshastalentedacquisitionseditors,andLisaMcClainis
oneofthebestinthebusiness.Sheheraldedthiseffortinthelatterstages,andit
hasbeenaprivilegetohaveservedwithheronthisproject.
Nowlet’stalkaboutindexing.Indexingabookisacrucialjob.Thisistrue
whetherthebookisinprintedordigitalform.Ijudgetechnicalbooksbyafew
keyindicators,includingthequalityandrelevanceoftheindex.Igenerally
alreadyhaveanideaofwhatI’mlookingforinatechnicalbook,sowhenIpick
upapapertechnicalbookforthefirsttime,Ifliptotheindexandcheckfor
particularsubjectareasofinteresttodeterminewhethertheindexcorrectly
referencestheminthenarrative.Ifyes,Ibuythebook.Ifnot,Idon’t.That’s
oftenmyonlycriteriawhenevaluatingnewbooks.That’showimportantthe
indexofatechnicalbookistome.
Ifyou’rereadingthisdigitally,youmightbethinking,“Sowhat?Icanalways
doawordsearch.WhydoIneedanindex?”Therearemanytopicsinbookslike
thiswherethekeywordisn’teasilysearchable.Forexample,considertheword
set.Itisusedquiteabitinnormalconversationandislikelyinthenarrativefor
variousreasons.ButsetisalsoakeyconceptinthisbookwithregardtoSET
operationssuchasUNIONandINTERSECT.Andthattopiciscompletely
differentfromthekeywordSETinanUPDATEstatement.Allareinthisbook.
Soifyouaresearchingforjustoneparticularuseofsetinthisbook,thenan
electronicsearchonakeywordmayslogyouthroughallsortsofunrelated
material.Butawell-craftedindexwillgetyouwhereyouneedtogoquickly,
eveninadigitalversionofthebook.
Forthatreason,IamverygratefulfortheworkofJackLewis,theindexerfor
thisbook.It’sagreatdealofworkandoftenoverlookedbutneverbyme—
neitherasaconsumernorasanauthor.Thankyou,Jack.Iappreciateyourwork
muchmorethanyoumayrealize.
Ahugeandspecialthank-youtomytechnicaleditor,ToddMeister.Toddhas
workedtirelesslytoreviewandanalyzehundredsofpagesofcode,andhehas
reverse-engineeredamyriadofexamples,illustrations,questions,andanswers.
Hetraveleddownthislongroadwiththerestoftheteamwithstyle,grace,and
amazingpersistence.Oneparticularchallengewiththiskindoftechnicalbookis
thatitdoesn’tlenditselftoaconsistentlybuiltapplication(forexamplesand
illustrations)likesomeotherbooksdo.Thenatureofacertificationexamguide
isthatitmustcoveravarietyofsyntaxcombinations,soratherthanbuildingan
applicationasaproject,thisbookrequiresyoutounderstandallformsofcertain
syntaxcombinations.Thatisachallengetothecertificationcandidate;imagine
thechallengetoatechnicaleditor.Itisenormous.YetToddhasdoneit
brilliantlyandwithaplombandprofessionalism.Iwillalwayshavevisionsof
hisfamiliar“ToddMeisterverified”annotationsthroughoutthedraftcopiesof
thebookasthevarioussectionspassedbackandforthalongthecycleof
production.Butnotethatthebookhasgonethroughafewiterations,soifyou
stillbizarrelymanagetofindsomethinginitthatislessthanideal—andthat’s
thecasewithvirtuallyeverypublication—thenthat’llbeonme.Butknowthat
thankstoTodd,thebookislight-yearsaheadofwhereitotherwisewouldhave
been.
Alsoabigthank-youtoBobBryla,anoutstandingOraclePressauthorinhis
ownrightmanytimesover,whojumpedinatthelastminutetocontributea
goodnumberofquestionsandanswerstothesamplecertificationexamincluded
withthebook.Thankyou,Bob!
TomyverymanyfriendsandcolleagueswithwhomI’vehadthepleasureof
workingand/orservingovertheyearsatvariouslocationsandinvarious
capacities,atsuchenterprisesasAtrexa,LHRDigitalConsulting,Corbinian,
Sysorex,ISC,db-Training,MAOP,EOUG,Boeing,ORI,ARC,NPC,and
elsewhere,includingsomewhoarestillatthoseplacesandsomewhohavegone
ontootheradventures—thereisnowayIcouldnameeveryoneherewhohas
beeninstrumentalorcontributedsomethingimportanttomylifeandwork.A
partiallistincludesthefollowing:MarkO’Donnell,MattHutchinson,Daveand
JackieDougherty,JohnCook,TimGreen,JeremyJudson,SalamQureishi,Nadir
Ali,WendyLoundermon,AtharJavaid,DanDoherty,TonyCovert,Bianca
Canales,MarleneTheriault,KevinLoney,MichelleVeghte,ArdellFleeson,
MarkusDale,KarlDavis,AllenShoulders,EdWolfe,AshleyRubeck,Cindy
Shiflett,PhilHasse,DaveGagner,JonFeld,JayNielsen,SteveSmith,Edgar
Kline,DaveSalmen,OscarWood,JamesPerry,TerriBuckler,SarahHenrikson,
MarkTash,SteveVandivier,AdhilShaikh,MoniqueTribolet,EdSpinella,Dino
Merkezas,KathyGardos,BertSpencer,KarenOwens,MikeAult,Graham
Seibert,VinceAdams,BobSchnetter,JoshParreco,CraigKasold,JenniferBlair,
BrettSimpson,MikeGerrity,DaveCooper,TedCohen,SteveCummings,
JimmySmith,PeterDube,RuthieWashburn,KimCurl,ToniBoyd,RobinRuth,
ReneeBattle,DannyDuong,HungNguyen,DrewDaffron,KenO’Neal,Kim
Miller,JohnLauder,BobSmout,ToddStottlemeyer,PaulLeslie,DavidWise,
DanRutherford,LauraTaylor,LauraSetliff,TrinTranh,WilsonDizard,Paul
Elliott,JohnMetelsky,DonKnight,ArtGarrison,MarshallCohen,MarkWojno,
BillMcCarren,JonathanSalant,CaroleFeldman,TammyLytle,SheilaCherry,
RickDunham,DougHarbrecht,AudreyFord,TimAquilino,DebbieBeebe,Bill
Simpson,AnnetteTaylor,FredWills,CarleszaHarris,GardnerMcBride,Cindy
McBride,JimFlyzik,RobGuerra,JohnCoffey,LyleBeall,BobbieBeall,andto
RoyPatterson,who–withJimBauchspies-assignedmetomyfirstOracle
project.AndtoToddandCindyBauchspies,MikeandKateWatersandGavin,
JamesandClaudiaWaters,PhilandCharlotteJones,HarrietMarinandJoe
Motz,andDannaHenderson.
Tothosewhoarenolongerwithus:Aaron“Eppie”Epstein,MartinKuhn,
JohnCosgrove,MikeShanahan,GordonGustin,HelenKamajian,Georgine
Bauchspies,EileenWaters,andJackHenderson.
ToDanandBrendaHinkle,whogavemeoutstandingopportunitiesandfrom
whomIlearnedagreatdealaboutOracleandentrepreneurship.
TomyverymanyoutstandingOraclestudentsovertheyears,toonumerousto
mentionhere.Irememberoneclassthatgavemeagoing-awaypresentofa
smallantiquewaterpump,symbolicforhowlearningisasimportanttoone’s
lifeasiswaternecessaryforsurvival.Thatsymbolicgesturewasmoving
beyondwords,andIhaveneverforgottenit.
TomylongtimeprofessionalassociateandclosefriendBillBryant,whohas
mentoredandcoachedmeonmanyaspectsoftechnology,business,theworld,
andlife.
Tomyfather,Don,abrilliantengineer,excellentprojectmanager,published
author,greatprofessional,andgreatdadandtowhomIdedicatedmysecond
book.
Tomymother,Joan,thebestmotherintheworld,whoIlovewithallmy
heartandtowhomIdedicatedmyfirstbook.
Averyspecialthank-youtoLisa,theloveofmylife,forbeingwonderful,
sweet,patient,brilliant,soloving,andsoveryencouraging!
Andfinally,toJimBauchspies,whoislikeasecondfathertomeinmany
ways,andtohislatewife,Georgine,whoseinfectiouslaughandwarmthand
lovearetreasuredmemoriesIwillneverforget.JimgavememyfirstOracle
opportunity,aswellasagreatdealofadviceonallthingspersonaland
professionaloverthemanyyearsI’vebeenprivilegedtoknowhim.ItistohimI
dedicatethis,mythirdbook.
R
PREFACE
ecentlyIattendedatechnicalconferencehostedbyaleadingbigdatavendor.
Ifoundmyselfspeakingwithanexhibitorabouthisstartupcompany’s
NoSQLproduct.Nowfolks,Iloveallformsofdatasystems;theyallhavetheir
place.Key-valuepairs,dynamicschemas—theyallhaveause.Butthis
particularvendorboldlydeclaredthattherelationaldatabasewasuselessand
“SQLisdead.”Hedescribedhowhissystemcouldharvesttext-basedcomments
andinfersentimentacrossmillionsofrecordsandsupportallsortsofcleveruse
cases—allofwhich,incidentally,isdoablewithOracle-drivenSQL-based
systems.ButhewasconvincedthatOracle’srelationaldatabasemanagement
system(RDBMS)wasirrelevant.
SoIaskedhim,“Let’ssayIplaceanorderforthisproductofyours.Areyou
goingtotrackthatorderinaNoSQLdatabase?Ordoesyourofficerelyonan
RDBMStocaptureorders?”
Hethoughtaboutit.“Goodpoint,thatisarelationaldatabasesystem.”
“OK,andhowaboutyouremployees?Whenyourcompanytracksyour
hours,whatdotheyuse?”
“True,true,that’saSQLdatabasetoo.”
“Howaboutsupplymanagement?Shipping?Generalledgerandother
accountingfunctions?Inventorymanagement?Andtheotherofficesystems?”
“OK,yougotme—thosearealsoRDBMSservers,allSQLsystems,true.”
“Soanythinginvolvingorders,customers,employees,vendors,partners,
products,inventory,manufacturing,payroll,andjustabouteveryotheraspectof
thebusinessisaSQLsystem.”
“True,it’strue.”
“ButifIwanttoanalyzecommentsonmywebsite,yourcompany’sproduct
isagoodchoice.”
“Yes!Exactly!”Heseemedrelievedtobebackonfamiliarterritory.“Andwe
cananalyzemillionsofcommentspersecond.”
“OK,soifIgetmillionsofcommentspersecondonmywebsite,I’dbe
interested.”
“Yes.Twitterusesus.”
“Cool.SoifIwanttoperformsometextanalysisonamassivescale,Icould
theoreticallystilluseOraclebutyourproductmightbebetteroptimizedfor
massivelylargevolumesofinsert-onlyprocessingifI’mTwitter.”
“Exactly.”
“Awesome.Oneproblem.I’mnotTwitter.HowmanyTwittersarethereout
there?LastIcheckedtherewasone.Nowgranted,Iknowotherorganizations
haverequirementstoprocessmillionsofrecordspersecond.Butwho?
Facebook?OrGoogle?Soperhapstwo,ten,orthirtywebsites?”
“Well…probablymorethanthat.”
Sure,probably.Andweallknowthatnumberisgrowing.Butstill—to
suggestthismeansthedeathofOracle?
AndatthispointIdecidedtoreallyfreakhimout.“Yourproductisprobably
writteninJava,isn’tit?”
“Yes,itis.”
“YouknowOracleownsJava,right?”Helookedhorrifiedanddeniedthatit
wastrue.
Folks,I’masexcitedaboutthebigdatarevolutionandthepotentialofthe
InternetofThings(IoT),machinelearning,andalltechnology.Iworkinthe
worldofbigdatasystemsandamyriadofexoticdatamodelingalternatives.I
lovecutting-edgedevelopments.Butattheendoftheday,thereisnoquestion
thattheworkhorseofallindustryandgovernment,thebackboneofautomated
systems,thetechnologyonwhichvirtuallyalllargeorganizationsrelyhasbeen
—andcontinuestobe—therelationaldatabasemanagementsystem.Andthe
OracleRDBMSisthebestinthebusiness.
IrecentlyattendedatechnicalsessionatapopularJavaconferencewherethe
speakerpresentedsomethinglike32usecasesforvarioussituationswheresome
newNoSQLsystemwouldperformfasterorotherwisemoreoptimallythan
Oracle’sRDBMS.Eachexamplewasbasedonauniqueand—tome—unusual
(perhapssomewouldsaycontrived)usecase.“Soifyouhavethisunusual
combinationofdatastructuresandyourbusinessobjectiverequiressome
particulartypeofanalytics,sure,youcoulduseOracle,butifyouwantashorter
developmentcycleorfasterperformance,youshouldusethisNewNoSQLOpen
SourceTool#17.”Foroneexample,theSQLcoderequiredwas30linesin
length,butdoingthesamejobwiththealternativenewtoolrequiredonly23
linesofcode.OK,sure,soI’mgoingtolearn,install,configure,code,test,and
deployadifferenttooleachtimeforeveryoneof32differentusecasessoIcan
typesevenfewerlinesofcode.Yeah,let’sdothat.
Ornot.HowaboutIjustuseOracleandbedonewithit?
Again,folks,don’tgetmewrong—allsystemshavetheirplace.Buteven
today,themostpowerfulinformationtoolintheworldisstilltherelational
database.Asofthiswriting,theofficialrankingatDB-ENGINES.comliststhe
topfourdatabaseenginesinusetodayasbeingoftherelationaldatabasemodel,
withsevenofthetopteninusefallingintothatcategory(seehttp://db-
engines.com/en/ranking).Atthetopofthatlististhesamedatabasethathasled
theworldforyearsnow:Oracle’sRDBMS.
Oracle’sMySQLisnumbertwo.
ThecorelanguageatthefoundationofallOracleproductsisOracle’s
StructuredQueryLanguage(SQL),alanguagethatiscommontoallmajor
relationaldatabasesofallvendorsworldwide.
Evenmanyoftheneweralternativedatastoresthataren’tnecessarilybased
onarelationaldatabasemodelarestrugglingtoimplementSQL-likelanguages,
suchasCassandra’sCQLorIBM’s“SQL-likequeryinglanguageHiveQL”
(https://www.ibm.com/developerworks/library/bd-sqltohadoop2/)forusewith
Hadoopsystems.
Therelationaldatabasemodelisstillthenumber-onedatabasemodelinthe
world,andSQListheonlylanguagecommontoallrelationaldatabasemodels.
It’salsothestandardagainstwhichnewspecial-purposedatabasemodelspattern
newlanguagefeatures.
Thisbookisdesignedtohelpyoubecomecertifiedinthissignificant,
powerful,importantlanguageatthecoreofthesuccessfuloperationsof
governmentsandindustryworldwide.
YouhaveamadeabrilliantchoicetostudyandworkwithSQLandtouse
thisbooktoworktowardcertification.
Welcome!
InThisBook
Thisbookisorganizedinsuchawayastoserveasanin-depthreviewforthe
OracleDatabaseSQLexamforbothexperiencedOracleprofessionalsand
newcomerstoSQLtechnologies.Eachchaptercoversamajoraspectofthe
exam,withanemphasisonthe“why”aswellasthe“howto”ofworkingwith
andsupportingrelationaldatabaseapplications.
DigitalContent
Formoreinformationregardingthedigitalcontent,pleaseseetheAppendix,
“AbouttheDownload.”
ExamReadinessChecklist
AttheendoftheintroductionyouwillfindanExamReadinessChecklist.This
tablehasbeenconstructedtoallowyoutocross-referencetheofficialexam
objectiveswiththeobjectivesastheyarepresentedandcoveredinthisbook.
Thechecklistalsoallowsyoutogaugeyourlevelofexpertiseoneachobjective
attheoutsetofyourstudies.Thisshouldallowyoutocheckyourprogressand
makesureyouspendthetimeyouneedonmoredifficultorunfamiliarsections.
InEveryChapter
I’vecreatedasetofchaptercomponentsthatcallyourattentiontoimportant
items,reinforcevitalpoints,andprovidehelpfulexam-takinghints.Takealook
atwhatyou’llfindineverychapter:
EverychapterbeginswithCertificationObjectives—whatyou
needtoknowtopassthesectionontheexamdealingwiththechapter
topic.Theobjectiveheadingsidentifytheobjectiveswithinthechapter,
soyou’llalwaysknowanobjectivewhenyouseeit!
ExamWatchnotescallattentiontoinformationabout,andpotential
pitfallsin,theexam.Thesehelpfulhintsarewrittenbytheauthor,who
hastakentheexamandreceivedhiscertification—whobettertotellyou
whattoworryabout?Heknowswhatyou’reabouttogothrough!
OntheJobnotesdescribetheissuesthatcomeupmostoftenin
real-worldsettings.Theyprovideavaluableperspectiveoncertification-
andproduct-relatedtopics.Theypointoutcommonmistakesandaddress
questionsthathavearisenfromon-the-jobdiscussionsandexperience.
TheCertificationSummaryisasuccinctreviewofthechapterand
arestatementofsalientpointsregardingtheexam.
✓ TheTwo-MinuteDrillattheendofeverychapterisachecklistofthemain
pointsofthechapter.Itcanbeusedforlast-minutereview.
Q&A
TheSelfTestoffersquestionsintendedtocheckyourknowledgeof
eachchapter.Theanswerstothesequestions,aswellasexplanationsof
theanswers,canbefoundattheendofeachchapter.BytakingtheSelf
Testaftercompletingeachchapter,you’llreinforcewhatyou’velearned
fromthatchapter.
SomePointers
Onceyou’vefinishedreadingthisbook,setasidesometimetodoathorough
review.Youmightwanttoreturntothebookseveraltimesandmakeuseofall
themethodsitoffersforreviewingthematerial.
1.RereadalltheCertificationSummaryandTwo-MinuteDrillsections,
orhavesomeonequizyou.Youalsocanusethedrillsasawaytodoa
quickcrambeforetheexam.Youmightwanttomakesomeflashcardsout
of3×5indexcardsthathavetheTwo-MinuteDrillmaterialonthem.
2.Re-readalltheExamWatchnotes.Rememberthatthesenotesare
writtenbytheauthor,whohastakentheexamandpassed.Heknowswhat
youshouldexpect—andwhatyoushouldbeonthelookoutfor.
3.Re-taketheSelfTests.TakingtheSelfTestsrightafteryou’vereadthe
chapterisagoodideabecausethequestionshelpreinforcewhatyou’ve
justlearned.However,it’sanevenbetterideatogobacklateranddoall
thequestionsinthebookinonesitting.Pretendthatyou’retakingthelive
exam.Whenyougothroughthequestionsthefirsttime,youshouldmark
youranswersonaseparatepieceofpaper.Thatway,youcanrunthrough
thequestionsasmanytimesasyouneeduntilyoufeelcomfortablewith
thematerial.
4.Takethepracticeexams,timedandwithoutyourbook,toseehowyou
did.Makenotesasyouprogresstokeepalistoftopicsyouthinkyouneed
tostudyfurtherbeforetakingtherealexam.(SeeAppendixAformore
informationaboutthepracticeexams.)
T
INTRODUCTION
hisbookhelpsyoubecomecertifiedintheStructuredQueryLanguageby
preparingyoufortheOCAOracleDatabaseSQLCertifiedAssociateexam,
1Z0-071.Theobjectiveofthisstudyguideistoprepareyouforthe1Z0-071
exambyfamiliarizingyouwiththetechnologyandbodyofknowledgetestedon
theexam.Becausetheprimaryfocusofthebookistohelpyoupassthetest,I
don’talwayscovereveryaspectofSQL.Someaspectsofthetechnologyare
coveredonlytotheextentnecessarytohelpyouunderstandwhatyouneedto
knowtopasstheexam,butIhopethisbookwillserveyouasavaluable
professionalresourceafteryourexam.
Chapter1startswithasectionthatwillprovideyouwithsomeintroductory
informationabouttheexamandtheexamexperience.TherestofChapter1
delvesrightintothetopicsofthecertificationobjectives.
Thebookalignstothecertificationexam’sstatedobjectives,publishedby
OracleCorporationandstatedbythecompanytobethetopicsfromwhichthe
examdrawsitsquestions.Thisbookwillnot—andcannot—revealactualexam
questionsortheircorrespondinganswers.Thoseofuswhohavetakentheexam
areforbiddentorevealthatinformation.Butthisbookwillfocusonthetopics
thattheexamaddressesandwillteachyoutheinformationthat,asofthis
writing,youneedtoknowabouttheexam.
Thisbookiscomprehensiveandfocused.NotethatOracleCorporationhas
madeseveralmanualsavailableonlinethatdescribethefullfunctionalityofits
products.Butthoseofuswhohavebeeninthebusinessawhileknowthatwhile
thesearevaluableresources,theyarehuge—somearethousandsofpageslong
—andsometimestheyareoverwhelmingtonavigateeffectivelywithina
person’sprofessionallifetimewithoutsomesortofguide.Thisbookisyour
guide,butitismuchmorethanthat;itisaself-containedstudyguide,complete
withfulldescriptions,syntaxdetails,samplecode,self-tests,andpracticeexams
thatsimulatetherealcertificationexamexperience.Inotherwords,youare
holdingatreasuremaptothenuggetsofwisdomyouneedtoknowtopassthe
examandwinyourcertification.Itistheproductofyearsofexperience,earned
throughhardwork,testedamongveteranOracleprofessionalsfromaroundthe
worldandwithmanybackgroundsandstrengths,andconsolidatedintoone
clearlyorganizedformattoempoweryoutopreparequicklyandefficientlyto
becomecertified.
Thebookisdesignedtoservethefollowingaudiences:
Fortheveteranwhowantstozeroinontopicsonanàlacartebasis,
thebookiscategorizedbycertificationobjective.Ifyouhavealready
seenthepublishedcertificationobjectivesandwanttostudyuponafew
areas,youcandothatwiththisbook—findtheappropriatetopicand
studythosechapters.
Forthereaderwhowantsamorecomprehensivereview,the
objectivesandchaptersaresequentiallyorderedtobeginwiththe
fundamentalsandworkuptothemoreadvancedtopics.Youcanstudy
thebookstraightthroughandexperienceacompletepresentationofthe
knowledgeyouneedtopasstheexam.
Fortheseasonedpractitionerwhowantstojumpstraighttotheexam
experience,gototheonlinepracticeexamsandtakeanexam.(Seethe
Appendixonhowtoaccessthepracticeexams.)
The1Z0-071examhasbeenvalidatedagainstOracleDatabaseversions11g
Release2(version11.2.0.1.0)andupto12cRelease1(12.1.0.1.0).Forthis
book,IusedOracleDatabase12cRelease1.Forthescreenshots,Iusedscreen
capturesoftwareto“grab”imagesofSQLstatementsasdisplayedineither
SQL*PlusorSQLDeveloper.NotethatintheSQLDeveloper’sScriptOutput
display,numericdatadisplaysleftjustifiedbydefault,asopposedtoSQL*Plus,
wherenumericdatadisplaysrightjustifiedbydefault.Fortheentity-relationship
diagrams,IusedOracleSQLDataModeler.
Goodreading,studywell,andthebestoflucktoyou.Yourfeedbackis
invitedandwelcome:soh@corbinian.com.
Exam1Z0-071
1
OracleandStructuredQueryLanguage
(SQL)
O
CERTIFICATIONOBJECTIVES
1.01TheExam:AnOverview
1.02IdentifytheConnectionBetweenanERDandaRelationalDatabase
1.03ExplaintheRelationshipBetweenaDatabaseandSQL
1.04DescribethePurposeofDDL
1.05DescribethePurposeofDML
1.06BuildaSELECTStatementtoRetrieveDatafromanOracleDatabase
Table
✓Two-MinuteDrill
Q&ASelfTest
racleCorporation’simplementationoftheStructuredQueryLanguage(SQL)
isarguablythemostpowerfulandmostsignificantcomputerlanguageused
intheworldofgovernmentandbusinesstoday.Thischapterbeginstheprocess
ofpreparingyoutosuccessfullytakeandpasstheOracle1Z0-071exam,Oracle
DatabaseSQLCertifiedAssociate.First,I’lldiscussafewparticularsaboutthe
examandhowitisdifferentfromotherOraclecertificationtests.You’lllearn
whatyoucanexpectonexamday.Next,I’llmakesomesuggestionsaboutstudy
materials.ThenI’llusetherestofthechaptertoanalyzethefirstexamobjective,
forwhichthischapterisnamed.Inthefirstofthesubobjectives,I’lldiscussthe
conceptsofarelationaldatabasemanagementsystem(RDBMS)andexplain
entity-relationshipdiagrams(ERDs).NextI’lltalkabouttherelationship
betweentheRDBMSandSQL,andyou’lllookcloselyattwoimportantsubsets
ofSQL:DataDefinitionLanguage(DDL)andDataManipulationLanguage
(DML).Finally,youwilllearnhowtobuildaSQLSELECTstatement.Thiswill
completethefirstobjectiveandsetthestagefortheremainderofthisbookas
youstudyeachsubsequentobjectiveandprepareforthetest.
IfyouareaveteranOracle-certifiedprofessionalwhoisalreadyfamiliarwith
theexamprocess,youmightskipthenextsectionabouttheexam.Infact,you
maychoosetojumptoonlythoseexamobjectiveswhereyouthinkyouneed
study.However,Iencourageyoutoreviewallthematerialinthisbook,evenif
youthinkyouknowitallalready.Theexamwillchallengeyouonfinepoints
andnuances,anditwilltrytotrickyouwithconceptsandkeywordsthatalmost
soundcorrectbutarenot.Youwantyourknowledgeofthematerialtoberock-
solid.Forthepurposeoftheexamaswellastothebenefitofyourcareer,itis
crucialthatyoumaintainyourSQLskillsatthehighestlevel,asthisiswhat
organizationsintheworldtodayrequire—andtheexamdemandsthesamefrom
you.Beconfidentinyourknowledgeandyou’lldramaticallyincreaseyourodds
ofsuccessonexamdayandinyourcareer.
Let’sgetstarted!
CERTIFICATIONOBJECTIVE1.01
TheExam:AnOverview
TheOracleDatabaseSQLCertifiedAssociateisademandingexam.Itposes
questionsthattestthefullbreadthofyourknowledgeofSQLsyntaxand
processinganditsapplicationtobusinessrules.Atypicalquestionontheexam
mightgosomethinglikethis:
Youmaybeaskedtoreviewanexhibit,whichcouldbeasetofdata
outputinahalf-dozencolumnsandperhaps20or30rows—oritmight
beanentity-relationshipdiagram(ERD)containingasmanyasahalf-
dozenentitiesormore.
YoumaybeaskedtoreviewarelatedsetofSQLstatementsthatare
intendedtooperateontheexhibityouwerejustshown,withanumberof
SQLstatementsinwhichtheremightbeaseriesofnestedscalar
functions,aggregatefunctions,subqueriesofvariousforms,andtheuse
ofdifferentstatementsandclausesshowcasingfeaturessuchasvery
largedatatypes,complexjoinconditions,andsoon.
Someofthecodemaybecorrect,andsomemaynot.You’llneedto
recognizethedifference.
WiththesampleexhibitandSQLcodeinfrontofyou,youmaybe
askedtoidentifytheresultingstatusofthedatabaseaftertheSQL
statementsexecute.
YoumaybeaskedtoidentifytheinternalworkingsoftheOracle
database,inorder,inaccordancewiththeSQLstatementsyou’vebeen
shown.
Thelistofpossibleanswersmayincludemorethanonecorrect
response,andyoumustidentifyeachofthem.
Doesthatsoundlikealottodoforasinglequestion?Thenconsiderthis:for
theentireexam,youareallowed100minutestoanswer73questions.Thatisan
averageof1.36minutesperquestion,orabout82secondseach.
LikemanyofOracle’scertificationexams,the1Z0-071examincludes
manymultiple-choicequestionsthathavemorethanonecorrectanswer.
Inotherwords,asinglequestionmayrequireyouto“selectallofthe
followinganswersthatarecorrect.”Theresult:youmustevaluateeach
individualanswer.Onceyoufindacorrectanswer,youaren’tdone—you
mustcontinuecheckingthemall.Onthesetypesofquestions,youcan’t
ruleoutanyoftheanswersbasedontheprocessofeliminationsinceall
answersarepossiblecandidatesforbeingcorrect.Onequestioninthis
formatismorelikeseveralquestionsrolledintoone.Theresultisamore
demandingtestthatrequiresyoutobemoreknowledgeableandrequires
moreofyourtimetoanswer.Sostudywell.Reviewthisbookthoroughly.
Thisisnotasimpleexam.Butyoucandoit—equippedwiththisbook,
youroddsofsuccessincreasedramatically.
Thinkyoucanhandleit?Doyouhavewhatittakestobeaformally
recognizedandofficiallycertifiedOracleDatabaseSQLCertifiedAssociate?
Whetheryoudoornotremainstobeseen,butonethingisforcertain.This
bookwillprepareyou,strengthenyourknowledge,fillinthegaps,and
dramaticallyincreaseyouroddsofsuccess.
Sogetreadyforafunandrewardingchallengeandanimportantmilestonein
yourcareer.Getreadytoentertheworldofthetechnicalelite,tojointhecrème
delacrème,andtoberankedwiththebestofthebest!
WhattoDoandWhattoExpect
Hereishowtoregisterfortheexamandwhatyoucanexpectasyougothrough
theprocess.
TestLogistics
Hereareinstructionsforhowtoregisterandtaketheexam.(Pleasenotethat
Oracle’sregistrationprocessissubjecttochange.SomeofthespecificsI
describeheremayhavechangedbythetimeyoureadthis.)
Toregisterforanexam,visittheOracleCorporationwebsite
(www.oracle.com),clicktheCertificationlink,andlookforthe1Z0-071exam
page.Fromthere,clickthelinkaskingtoregisterfortheexam.Yourbrowser
willnavigatetothewebsiteforthecompanythatproctorstheexam,Pearson
Vue,atwww.pearsonvue.com.Fromthatwebsite,youcanlocatealocalfacility
atwhichyoucantaketheexam.Lookforanavailabletime—oddsareyou’ll
havetoplantotaketheexamatleastonedayinadvance,ifnotlonger,perhapsa
week.Provideyourcreditcardinformationfortherequiredpayment,andyou’ll
bebookedtotaketheexamattheprecisetimeandlocationyouselected.
Whenyouarriveattheexamlocation,youmaybeaskedtoturnoffyour
mobilephoneandgiveittoastaffmember.Inmycase,thestaffmemberlocked
mymobilephoneinaportableheavy-dutybag.Thestaffretainedthekeytothe
bagbuthandedthebagcontainingmymobilephonetome.IwastoldIwouldbe
abletotakethebagwithmeintothetestingroom,unabletoaccessitinsidethe
lockedcontainer.IwastoldIcouldrecovermyphoneaftertheexam.
You’llneedtoprovidetwoformsofIDtothestafftocompleteyoursign-in
process.Afterthat,you’llprobablybeescortedintoaroomwithadeskwitha
computer,loggedintoanautomatedtestingsystem.Ifyouwanttotakenotesfor
useduringtheexamonly,youmaybeprovidedwithafewblanksheetsofpaper
andapenoradry-eraseboard.(Notethatwhatevernotesyoumighttake,you
willberequiredtoturntheminattheendoftheexam.)
Oddsareyou’llbeleftaloneatthatpoint,butnotethattheexamhasnotyet
started.Onthecomputer’sautomatedtestingsystem,you’llbepresentedwitha
seriesofdisclosuresandagreementstoreview,andwhenyou’reready,youcan
clickabuttontoinitiatethetimedexam.Thetimerwillstartwiththefirst
question.Atimeronthescreenwilltrackyourremainingtimeasyouprogress
throughtheexam.
Allthequestionswillbemultiple-choicequestions.Manyofthequestions
willrequireyoutoclickabuttontodisplayanexhibit.Theexhibit—once
displayedbyyou—willprobablypopupinaseparatewindow,automatically
sizedjustbigenoughtoshowwhatevertheexhibitisdisplaying.Theexhibit
maybeanentity-relationshipdiagram,oritcouldbealistingofdatathatcould
havebeenthecontentsofatableortheoutputofareport.Manyquestions
includeSQLcode.
Duringmyexamexperiences,Ihavenoticedthatmanyquestionsareoften
accompaniedbyagoodnumberofdiagramsorcodeordataoutputnot
necessarilyrelevanttothepointofthequestion.Mostquestionswillpresentto
you,forexample,anERDandadatalisting,aswellasabriefexplanationof
somebusinessrulesorbackground,andfinallyasktheactualquestion.You’ll
eventuallyrealizethatyoucanwasteagoodamountofthetimeanalyzing
detailsofexhibitsbeforeyoufinallyreadthequestion,onlytorealizethatyou
didn’tneedtostudytheentireexhibitafterall.Manyquestions,onceunderstood,
canbeansweredataglanceoftheERD.Sodoyourselfafavor:whenyou
encounteranewquestion,skippastthecodeandthedataoutput(thereport,
exhibit,orwhatever)andskipaheadtotheactualquestion,whichisoftenthe
finalsentencejustbeforethemultiple-choiceanswersarepresented.Readthe
questionandthengobackandquicklyskimtheprecedingmaterial;you’llfind
thatyoucanidentifytheanswermuchmorequicklythiswaythanbyreadingall
thebackgroundmaterial.
Yourtimeislimited.Lookforthequestionfirst.Stayfocusedonthe
question.Don’tletalargeexampleofSQLtextdistractyou.Youhave73
questionstoanswerin100minutes.Thatis82secondsperquestion.
Somequestionswilltakeseconds;somewilltakemuchlonger.Watch
yourtime.
Thequestionswillbepresentedonthescreenoneatatime.ClickNextto
advancetothenextquestion.Youcanskipquestions;youaren’trequiredto
answereachquestionbeforeadvancing.Toflagaquestionforlaterreview,each
questiondisplayhasanoptionalMarkcheckboxintheupper-leftcorner,which
youcanuseto“mark”anyquestionforfuturereference,whetheryouhave
answereditornot.
Onceyouhaveprogressedthroughallthequestions,youwillreacha
summaryscreenshowingthenumbercorrespondingtoeveryquestionofthe
examinasingulartabularlisting.Thequestionswillbeidentifiedbynumber
only,andnexttoeachnumberwillbetheletter—orletters—oftheanswersyou
haveprovided.Anyquestion“marked”willbeindicatedbyahighlightedMnext
toit.AnyquestionnotfullyansweredshowswithahighlightedI—for
“incomplete”—nexttoit.You’llbeabletoeasilyreviewandcompletethe
answersandreviewanyquestions,includingthoseyouhavemarked,beforeyou
indicatethatyouhavecompletedtheexam.
Onenuancetokeepinmindoccurswhenyouencounteraquestionwithmore
thanonecorrectanswer.Toscorethosequestions,youwillhavetoprovidethe
correctnumberofcorrectanswers.Thequestionmaystate“choosethetwo
correctanswers,”forexample.Youwillhavetoprovidetwoanswers.Ifyoutry
toprovidethree,theautomatedsystemwillstopyou.Itwillpopupasmall
messagewindowtellingyoutofirstdeselectanexistingchoicetoensureyou
providenomorethanthecorrectnumberofanswerselections.
Butwhathappensifyouselecttoofew?Forexample,whathappenswhena
questionrequiresthreeanswersbutyouselectonlytwoandthenmoveontothe
nextquestion?Nothinginthesystemwillstopyoufromtakingthatstep.Ifyou
dothat,you’llbeleavingthequestionincompletewith—inthisexample—only
twoofthethreerequiredanswers.Youcanalwaysadvanceandleaveany
question“incomplete.”However,beforeyoucansubmittheexamforgrading,
youwillbeautomaticallypresentedwiththesummaryscreenattheend,in
whichanyincompletequestionswillbeflaggedclearlywithanIfor
“incomplete.”Youwillbeabletogobackandreview.
Reviewthesummaryscreencarefullyforanyunansweredquestions.
Lookforyour“marked”questionsthatmightrequirereview.Anddon’t
worry—youcan’tsubmittheexamforgradingwithoutfirstgoing
throughthesummaryscreen.Keepaneyeoutforit,andinspectit
carefully.
Whenyouaredonewiththequestionsandaresatisfiedthatyouhave
answeredeverything,clickExitonthesummaryscreen.Thetestscorewillbe
instantlyevaluated,andyoumaybeshownyourscoreandpassinggradeonthe
screen.Ifnot,you’llprobablybee-mailedtheresultswithin24hours.
Atthatpoint,pickupanypapersyoumayhaveorlockedbagscontaining
yourvaluablessuchasyourmobilephone,andvisitthefrontdeskofthetesting
center,whereaclerkmayprovideyouwithacopyofyourtestresults.Theclerk
canalsounlockthecontainercontainingyourmobilephone,whichyoucan
retrieve,andyoucanexitthefacilityknowingyouhavesatisfiedthe
requirementstobecomeahappyOracleDatabaseSQLCertifiedAssociate.
OracleSQLvs.ANSISQL
ThecertificationexamwilltestyouonOracleSQL.OracleSQLiscloseto,but
notidenticalto,thestandardestablishedforSQLbytheAmericanNational
StandardsInstitute(ANSI),alsoknownasANSI-standardSQL.Youwillnotbe
requiredtoknowthedifferencesbetweenthem.
OracleSQLvs.OracleSQL*Plus
ThecertificationexamwilltestyouonOracleSQLbutnotonOracle’s
enhancementstoSQLknownasSQL*Plus,withtwoexceptions:theDESC
commandandsubstitutionvariables.
SQL*Plusisasetofcommands,anditisalsoasoftwaretoolwithan
interfaceintowhichyoucantypeSQLandSQL*Pluscommandsandmonitor
theirexecution.SQL*Pluscommandsincludealargenumberofoptions,many
ofwhicharedevotedtoformattingoutput.Thosecommandsarenotincludedin
theexam.Youwon’tbestudyingSQL*Pluscommandsinthisbook,otherthan
DESCandsubstitutionvariables.
NotethatIwillusetheSQL*Pluscommand-lineinterfacefromtimetotime
todemonstrateOracleSQLcommands.
SQLFundamentalsIvs.SQLCertifiedAssociate
Sinceyou’replanningonobtainingyourcertificationbytakingthe1Z0-071
exam,thenchancesareyoumayhavealreadylookedat,oreventaken,the1Z0-
051exam,SQLFundamentalsI.Thetwoexamssharesomecommonobjectives,
but1Z0-071goesbeyond1Z0-051.SeeTable1-1foracomparisonofthose
objectives,includingwherethetwoexamsaresimilarandwheretheyare
different.Thetableliststheobjectivesfor1Z0-051ontheleftand1Z0-071on
therightandlinesuptherowstoshowwheretheexamssharecommon
objectivesandwheretheyvaryfromeachother.Herearesomeexamples:
1Z0-051testsontheCartesianproduct(6.4),but1Z0-071doesnot.
Bothexamstestonthe“RestrictingandSortingData”objectiveand
itsthreesubobjectives,but1Z0-071addstwoadditionalobjectivestoit,
coveringampersandsubstitutionandtheSQLrowlimitingclause.
Only1Z0-071testsoncorrelatedsubqueries,EXISTS,NOT
EXISTS,andWITH,aswellasthedatadictionary,controllinguser
access,multitableinserts,andMERGE.
Only1Z0-071testsontheobjectives“ManagingObjectsUsingData
DictionaryViews,”“ControllingUserAccess,”and“ManipulatingLarge
DataSets.”
ThosearejustsomeexamplesoftheinformationinTable1-1.Notethat
Oracledoesnotnumberitsobjectives,soItookthelibertyofassigningnumbers
totheobjectives,respectingthesequenceoftheoriginallypublishedlistat
Oracle’swebsite.
TABLE1-1 Comparisonofthe1Z0-051and1Z0-071ExamObjectives
Asyoucansee,theSQLCertifiedAssociateexamgoesfurtherthanthe
topicsaddressedbytheSQLFundamentalsIexam.
TheSQLFundamentalsIexamconsistsof64questionsin120minutes;SQL
Associateconsistsof73questionswithin100minutes.Atthetimeofthis
writing,thepassingscorerequirementspublishedattheOracle.comwebsiteare
60percentforSQLFundamentalsI
63percentforSQLAssociate
However,notethatpassingscorerequirementsaresubjecttochangewithout
notice.OracleCorporationreservestherighttosubstituteoneversionofthetest
withanotherversion,anddependingonthecomplexityofthespecificquestions
includedinanewversion,thepassingscoremaybeadjustedaccordingly.Oracle
publishesthisnoticeonitswebsiteregardingtherequiredpassingscoreforany
givenexam:
OracleroutinelypublishesnewversionsofitsOracleCertification
exams.Thepassingscoreforeachexamversionissetindependentlyto
maintainaconsistentscoringstandardacrossversions.Forexample,to
maintainaconsistentstandardforsuccess,anewversionofanexamwhich
containsmoredifficultquestionsthanthepriorversionwillbeassigneda
somewhatlowerrequiredpassingscore.Forthisreason,passingscores
mayvarybetweendifferentexamsinyourcertificationtrack,andbetween
laterandearlierversionsofthesameexam.Oracledoesnotrecommendan
exampreparationstrategytargetingthepassingscore,aspassingscoresare
subjecttochangewithoutnotice.
Inotherwords,studywell,anddon’tlimityourgoaltosimplyachievingthe
minimalpassingscorerequirement.Instead,dothebestyoupossiblycanto
increaseyourchancesofvictory.
SubjectAreas
Manyquestionschallengeyourknowledgeofseveralfactsatonce.Forexample,
Iencounteredonequestionthatpresentedseveralnestedscalarfunctionsina
seriesofSELECTstatements.Ihadtodothefollowing:
Understandclearlywhateachindividualscalarfunctiondid
Recognizesyntaxissuesthroughoutthecode
Understandthedatatypetransformationsasonefunctionpassedon
resultstoanother
Confirmwhethertheparameterpositioningwasaccurate
Identifytwofactsabouttheprocessandendresult,allwithinthe
contextofthegivenERD
Themoraltothestoryisthatyoushouldstudythisbookwell,understand
everythinglistedinthecertificationobjectives,andgetallofyourfactsdown
cold.Ontestday,showuprestedandontimeanddon’tgetdistracted.Witheach
question,seekouttheactualquestionfirstandthenlookattheexhibitsandcode
listings.Andwatchthetime.
You’llbegladyoudid.
ConfirmAppropriateMaterialsforStudy
Thissectiondiscussessomeitemsyoumaywanttogatherasyouprepareto
studyfortheexam.IfyouareaseasonedveteranintheSQLbusiness,youmay
notneedanyofit—thisbookwillsuffice.Butifyou’dliketoputforththatextra
bitofeffort,itmightbeagoodideatogetyoursoftwareanddocumentation
togetherasdescribedinthissection.
Software
OracleCorporationstatesthatithasvalidatedthe1Z0-071examquestions
againstOracleDatabase11gRelease2(11.2.0.1.0)andupto12cRelease1
(12.1.0.1.0).Iusedversion12c,Release1,inpreparingtheSQLstatementsfor
thisbook.
Ifyoudon’thavethesoftwareyouneed,youcandownloaditfromthe
officialOracleCorporationwebsite,atwww.oracle.com.
Ifyouhaven’tjoinedtheOracleTechnologyNetwork(OTN),thenyoushould
doitrightaway.Thereisnochargeforit.Visitwww.oracle.com/technetwork/
andsignuptoday.Fromthere,youcandownloadagreatdealofOracle
softwareforevaluationandstudy,includingthedatabaseitself.
IfyouinstallthepersonalversionofOracle,you’llprobablygetSQL*Plus
andSQLDeveloper,eitherofwhichyoucanuseforenteringandexecutingSQL
statements.Ifnot,thenbesuretodownloadoneortheothertoinstallanduse
withthedatabase.You’llneedoneorbothofthoseor,ifnotthem,thensome
sortoftoolforenteringandexecutingSQLcommands.Chancesareyoualready
havesomethingthatyou’reusingorelseyouprobablywouldn’tbeconsidering
certification.
Oracle’sToolsforWorkingwithSQL
MostofOracle’svariousproductsandtoolsuseSQL.ThisincludesOracle
productssuchasOracleFusionMiddleware,includingOracleSOASuite,as
wellasOracleAPEXandothers.Manydevelopmenttools,suchasOracle
JDeveloper,providetheabilitytoenterSQLstatementsandexecutethem.Two
ofthemostcommonlyusedtoolsforthispurposeareOracle’sSQL*Plus
command-lineinterfaceandOracleSQLDeveloper.
TheSQL*PlusCommand-LineInterfaceTheSQL*Pluscommand-line
interfaceisasimplewaytotypeSQLcommands,executethem,andobservethe
result.Itisauniversalsystemthatoperatesthesamewayineveryoperating
system.
SeeFigure1-1foranexampleofwhatthecommand-lineinterfacelookslike.
FIGURE1-1 TheSQL*Pluscommand-lineinterface
Theadvantagetothecommand-lineinterfaceisthatitfunctionsidenticallyin
theWindows,Unix,andLinuxoperatingsystems.Thatisoneofthemany
advantagesthatOraclehasalwaysoffered—easeofuseinanyoperatingsystem.
SQLDeveloperTheSQLDeveloperinterfaceisaninteractive,point-and-click,
menu-drivengraphicaluserinterface(GUI)thatispowerfulandgives
developersaquickoverviewoftheentiredatabase.Somecommandscanbe
enteredbyeithertypingthemorusingapoint-and-clickinteractionwitha
graphicmenu.SeeFigure1-2foranexampleofwhatSQLDeveloperlookslike.
FIGURE1-2 TheSQLDevelopertool
ThereareothertoolsthatprocessSQLstatements:
OracleJDeveloper
OracleApplicationExpress
SQLWorkshop
Others
Forthepurposesofthecertificationexam,yourchoiceofinterfaceis
irrelevant.SQLstatementsexecutecorrectlyinallOracleinterfaces.
TheexamwilltestyourknowledgeofthesyntaxofSQL,notyour
abilitytopointandclickyourwaythroughaGUI.Thefactthatyou
mightbeabletocreateaSQLtableusingacodegeneratorthrougha
point-and-clickinterfacewillnothelpyouduringtheexam.Ifyouarea
seriousapplicationsarchitect/programmer,therewillcomeatime—
probablyfrequently—whenyouneedtodesignorprogramafeatureas
partofalargerapplicationanddesignandembedSQLcodeintoother
programminglanguagesthathavenoaccesstotheniceGUItoolsduring
applicationruntime.Furthermore,asyou’llseeinnumerousinstances
inthisbook,therearemanytypesofSQLstatementsinwhichyoucan
combinefeaturesandclausesinsuchawaythattheyappeartobe
correct,executewithouterror,andproducelotsofoutput—allofwhich
canbetotallyerroneous.AtrainedeyeglancingattheSQLcodewill
recognizethemistake;anuntrainedeyewillnotevenrealizethereisa
problem.Inotherwords,thereisnoalternativetohavingcomprehensive
knowledgeofthesyntaxofSQL,neitherintheworldoftheserious
softwaredevelopernoronthecertificationexam.Knowyoursyntax.As
youstudyforthisexam,typeyourcommands,makesuretheyarecorrect,
andmakesureyouunderstandthemthoroughly.
Documentation
Thebookyouhaveinyourhandsisanoutstandingreferenceandisallyouneed
inthewayofdocumentation.Thisbookisthesinglebestguideyoucould
possiblygettoprepareyoufortakingtheexam.
Butifyoucravemore,youcandownloadadditionaldocumentationfromthe
OracleTechnologyNetworkwebsite,atwww.oracle.com/technetwork/.The
amountofdocumentationisalmostoverwhelming,particularlytoanewcomer.
Butonevolumeinparticularisofinterestforthepurposeofthecertification
exam:theSQLLanguageReferenceManual.Itisahugebook,withcloseto
2,000pages.ThesizeofthePDFversionismorethan14MB.Itssyntaxcharts
arecomplexandgobeyondtheneedsoftheexam.Thebookcontainsfarmore
informationthanwhatyou’llneedtopass,allofwhichisyetanotherreasonwhy
you’rebrillianttohaveobtainedthisbookyounowhaveinyourhands.Iwill
refertotheSQLLanguageReferenceManualfromtimetotime,butIwillfocus
onlyonthepartsthatarerelevanttopasstheexam.
OtherusefulreferencesincludeOracle’sAdvancedApplicationDeveloper’s
Guide,Concepts,SecurityGuide,GlobalizationSupport,andAdministrator’s
Guide.
Thequestionsforexam1Z0-071havebeentestedagainstthefollowing:
OracleDatabase11gRelease2
OracleDatabase12cRelease1
IwillrefertothesetofmanualsforOracleDatabase12c,Release1,whichis
alsoreferredtoas12.1.
CERTIFICATIONOBJECTIVE1.02
IdentifytheConnectionBetweenanERDanda
RelationalDatabase
Fortheremainderofthisbook,Iwillexamineeachoftheobjectivesoftheexam
itselfindetail.InthissectionIdiscusstheconceptofanentity-relationship
diagramandtalkabouthowitistransformedintoarelationaldatabase.Along
thewayI’lldiscussbusinessrequirements,datamodeling,theroleofprimary
andforeignkeys,andtheconceptofdatanormalization.
Entity-RelationshipDiagramsandDataModeling
Generallyspeaking,thereasonadatabaseexistsistosupportanexisting
businessprocessormodelsomeexistingreal-worldsystem.Forexample,a
typicalbusinesshasemployeeswhoservecustomersbyprovidingsomesortof
productorservice.Ifyouaretaskedwiththeresponsibilityofbuildingan
automatedbusinesssystem,youmightcreateadatamodelofthebusiness’s
operations.Tocreatethedatamodel,youidentifythebusiness’sbasic“things”
(orentities).Anentitymightbeacustomer,anemployee,oraproduct.You
mightaddmoreabstractconceptsasentities,suchasanorder,areturn,a
transaction,acustomer’sstorevisit,andsoon.
Asyouidentifyyourentities,you’llneedtodeterminetherelationships
amongthoseentitiesintermsofone-to-one,one-to-many,ormany-to-many
relationships.Forexample,you’llprobablyallowforanyonecustomertoplace
zero-to-manyorders.Whywouldyouhaveacustomerwithnoorders?It
dependsonyourbusinessrules—youmaywanttotrackpotentialcustomerswho
haveaskedforaprintcopyofyourcatalogbutperhapshaven’tyetordered
anything.
You’llprobablywanttoallowforanyoneordertoconsistofmultipleline
items,eachofwhichistakenfromyourmasterlistofavailableproducts.Each
one-lineitemmaybereturned—butnomorethanonceandideallynotatall.
SeeFigure1-3foranexampleentity-relationshipdiagramillustratingthis
initialdatamodel.
FIGURE1-3 Entity-relationshipdiagram
ThisexampleERDshowsfiveentitiesandtherelationshipsbetweenthem.
Eachentityisabox;eachrelationshipisalineconnectingtwoboxeswithsome
symbolstodetailthenatureoftherelationship.Forexample,anyonecustomer
willhavezero,one,ormanyorders.ThisisindicatedinbetweenCustomersand
Ordersbythelinethatshowsthecircle(zero)withthecrow’sfoot(many)onthe
sideofthelineconnectingtoOrders.Thecircleandcrow’sfootspecifythat
theremaybezero-to-manyordersforanygivencustomer.
ForanyonerowinOrders,theremaybeone-to-manyLineItemsentries,
indicatedbytheline(one)withthecrow’sfoot(many)neartheLineItemsentity
onthelinefromtheOrdersentity.
AnyoneProductsentrymaybeorderedzero-to-manytimesasaLineItem.
AndasingleLineItemsrowmaybereturnedzero-to-onetime,indicatedbythe
circle(zero)andbar(one)attheReturnsentity.
Onceyou’veestablishedtheinitialsetofentitiesandtheirrelationships,you
mightenhanceyourERDwiththeadditionof“attributes”toeachentity.See
Figure1-4.
FIGURE1-4 ERDwithattributes
Forexample,foreachcustomeryoumightwanttotrackthecustomer’sname
andcontactinformationsuchasane-mailaddressandmailingaddress.Foreach
order,you’llwanttocapturetheorderdateandperhapsthetermsofpayment.
Foreachlineitem,you’llwanttoknowthequantityorderedandperhapsa
discountfortheparticularlineitem,ifappropriate.
Let’sconsideranotherexample.ThefictionalCoddCruiseLinesisa
companythatprovidescruisevacationstoitscustomers.Thecompanyoperates
ships,andeachshipisassignedtoasinglehomeport.Eachshipisfilledwith
cabinsthatcustomersreservefortheircruisevacations.Thecompanyhas
employeeswhoareassignedtooneshipatatime.
Figure1-5isanexampleofwhatanERDforthiscruiselinemightlooklike.
FIGURE1-5 ERDforCoddCruiseLines,initial
Theexampleshowsthatforeachport,thereareoneormoreships,and
conversely,eachshipisassignedtooneandonlyonehomeport.Eachshiphas
zeroormoreemployees,suggestingthatyoumightstarttrackingshipsinyour
datamodelbeforetheyhaveemployeesassignedtothem—perhapswhilethe
shipisstillunderconstruction,inmaintenance,andnotinservice.Also,each
shipconsistsofoneormorecabins.
Ifyouweretostartassigningattributestotheentities,youmightendupwith
theexampleshowninFigure1-6.
FIGURE1-6 ERDforCoddCruiseLines,withattributes
Forexample,foreachshipyoumightwanttocapturetheship’ssizeand
capacity.Foreachemployeeyoumightwanttotracktheemployee’sname,start
date,andcurrentsalary.Whathappenswhentheemployeegetsaraise—doyou
retainarecordofthehistoricsalary?Notinthisexampledatamodel,butyou
couldextendthemodelbyaddinganentityfor“salaryhistory”orsomething
comparableandthenassignthatentitysomeattributes,suchassalaryamount,
initialdate,andfinaldateofserviceatthatsalary,andperhapsyoucouldeven
recordtheindividualauthorityintheorganizationwhograntedthatsalaryinthe
firstplace.
Thereisnosingleabsoluterightorwrongwaytomodelareal-worldbusiness
process.Theappropriatemodelwillvarybasedonwhatyouintendtocapture
andhowyouintendtoworkwiththesystem.Thefinalauthorityisthesetof
real-worldbusinessrequirementsyouaresupporting.Yourdatamodelisan
attempttorepresentthosereal-worldbusinessrequirements.
Forexample,inthefictionalCoddCruiseLines,youarecapturinghome
portstowhichshipsareassigned.Intherealworld,thereexistmanythousands
ofports.Areyouintendingtocapturethemall?No,notinthismodel,becauseit
isn’tnecessaryforthebusiness.Areyouintendingtocaptureeveryporttowhich
yourshipswilltravel?Notinthismodel.Youareintendingtocaptureonlythe
homeports.Butyoucouldmodifythemodelifyourbusinessrequirements
changedandyourintendedpurposeofyourautomatedsystemweretochange.
Thepurposeofadatamodelisnottorepresenteverypossibilitythatexistsinthe
realworldbuttocapturethatnecessarysubsettowhichtheautomatedsystem
willprovidesupport.
RelationalDatabases
Nowlet’sstarttouseanERDtocapturesomeactualdata.Rememberthatan
ERDisalogicalrepresentationofareal-worldsystem.Arelationaldatabaseisa
physicalimplementationofthatERDand,assuch,willneedtocaptureactual
data.Therelationaldatabasebuildsontheentitiesandtheirrelationshipsby
specifyingattributesoftheentities.Intherelationaldatabase,you’lladdprimary
keystouniquelyidentifyeachrowwithineachentity.Youmayalsoaddforeign
keystosomeentitiestorelatetherowsofthatentitytoanotherentity’srows.
Let’slookathowthisworks.ConsiderTable1-2,whichshowsalistofships
withthefictionalCoddCruiseLines.Thislistiswhatyoumightusetocreatea
databasetable.ThelisthascolumnsforShipIDandattributesoftheShips
entity.ThereasonyouneedaShipIDistomakesureyouhaveaunique
identifierforeachrowinthelist.Thefirstrowhasauniqueidentifierof1,the
secondrowhasauniqueidentifierof2,andsoon.Thisuniqueidentifierisakey
toidentifyingaparticularship’srow.ThevaluesfoundintheShipIDcolumn
uniquelyidentifyeachship.Thiscolumnisconsideredtobeaprimarykey
column.
TABLE1-2 CoddCruiseLinesShips
Now,let’ssayyoualsowanttocreatealistofemployees.You’llcreatea
primarykeycalledEmployeeIDandcapturealltheattributesofourEmployees
entity,alongwithsomethingadditional:ShipID.SeeTable1-3.
TABLE1-3 CoddCruiseLinesEmployees
Now,ifIweretoaskyoutoidentifytheshiptowhichMikeWestwas
assigned,whatwouldyousay?
Naturallyyouwould(orshould)sayitwastheCoddVictorious,andyou
woulddeterminethisbylookinginthelistofemployees,findingtherecordfor
MikeWest,“relating”thatrecord’sShipIDvaluetothelistofships,andfinding
thatShipID4“relates”totheshipnamedCoddVictorious.
ShipIDistheforeignkeyforEmployees.Aforeignkeyforagiventable
pointstotheprimarykeyofanothertable—inthiscase,Ships—toindicate
whichrowrelatestothecurrentrow.
Theselistsarewhatyouwillstoreindatabasetables.Andthedatacontained
withinthelistsareexamplesofthesortofdatathatarelationaldatabasemight
containintablesandthesortofprocessingitdoesto“relate”datainonetableto
datainanothertable.
Atypicaldatabaseconsistsofanynumberoftables,manyofwhichcontain
keyinformation(suchastheEmployeeIDandShipIDvaluesinthefigures)that
isusedtorelaterowsofonetabletorowsofanothertable.Themechanismused
toidentifyauniquerowwithinalistisaprimarykey,andyouuseaforeignkey
torelatedatafromonetabletoanother.
ThePRIMARYKEYisauniqueidentifierforarowinaparticular
table.
TheFOREIGNKEYisacopyofsome(orall)ofonetable’s
PRIMARYKEYdatainasecondtablesothatthesecondtablecanrelate
tothefirst.
Manyoftheexamquestionswillaskyoutoconsultan“exhibit”that
willfeatureanERDandthenaskyouaSQLquestionbasedonthetables
presentedintheERD.Thisbook’sselftestsincreasinglydothesameas
youprogressthroughthebook,andthepracticeexamsusethesame
approachastheofficialcertificationexam.
Many-to-ManyRelationships
Oneofthemostcommonchallengesintransformingalogicalmodelintoa
typicaltransactionalphysicalmodelhastodowiththesituationofamany-to-
manyrelationshipbetweentwoentities.Rememberthatthenatureofthe
relationshipbetweentwoentitiesisdefinedbytherequirementsofthebusiness
model.IntheearlierexampleofthefictionalCoddCruiseLines,thebusiness
requirementwasonlytotrackanemployee’scurrentlyassignedship.Butwhatif
therequirementweredifferent?Whatifyouneededtotrackallshipstowhichan
employeewereeligibletowork?Thiswouldsuggestthatanyoneemployee
couldbeassociatedwithzeroormoreships,andanyoneshipcouldbe
associatedwithzeroormoreemployees.
Whyzero-or-more?Itdependsonthebusinesslogic—let’ssayyoudecide
youwanttostarttrackingshipsbeforeemployeesareassignedandyoucapture
informationaboutemployeeswhoaren’tnecessarilyassignedtoships.Because
ofthosebusinessrules,you’llgowiththezero-to-manyrelationshipsinboth
directions.However,ifthebusinessdoesn’thavetheserequirements,thenyou
mightrequireone-to-manyrelationshipsinstead—itdependsonthebusiness
requirements.
Butregardlessofwhetherthesearezero-orone-to-manyineitherdirection,
thisparticularapproachtothebusinessresultsintwoentitieswithamany-to-
manyrelationshiptoeachother;seeFigure1-7.
FIGURE1-7 Many-to-manyrelationship
Theexampleshowsthattheships-to-employeesrelationshipismany-to-
many,meaningthatyouhaveone-to-manyrelationshipsinbothdirections:one
employeemaybeassignedtomultipleships,andoneshipmayhavemultiple
employeesassignedtoit.
Thepresenceofamany-to-manyrelationshipinanERDisamajorflagina
datamodel.Themessage:thedatamodelismissinganentity.Whileit’sfineto
showamany-to-manyrelationshipasyouareintheprocessofbuildingadraft
model,thereisneveramany-to-manyrelationshipinafinishedphysicalmodel.
Thereason:itdoesn’twork.Imaginetryingtotransformthisintothesortoflists
—ortables—youjustlookedatinthepreviousexample.Trytocreateyourown
lists—oneofships,anotherofemployees.Assignprimarykeyvaluestoeach
list.Thentrytofigureouthowyou’regoingtomaketheforeignkey
assignments.Youcouldtrytoaddoneforeignkeycolumntoonetableandthen
theother—evenboth.Butyou’lldriveyourselfcrazytryingtolistthedataso
thateachrowofshipsislistedonce,eachemployeeislistedonce,andthereare
foreignkeystorepresentthemultipleinterrelationshipsthatmightexist.Itwon’t
work.
Here’swhatworks:athirdentity.SeeFigure1-8.Thepresenceofamany-to-
manyrelationshipbetweentwoentitiesisalwaysanindicationthatathirdentity
belongsinbetween,andit’softenanentitythatismoreabstractinnature.Inthis
case,nametheentityRosteranduseittolisteachassignmentofanygiven
employeetoanygivenship,alongwithstartandstopdatesofwhenthe
employeeisassignedtoworkontheship.Thisenablesyoutolistasmany
repeatinglinesasrequiredforanygivenship—onceforeveryemployee
assigned—andtodothesamethingintheotherdirection,listingevery
employee’sassignmenttoshipaftership,asmanyasisrequired.
FIGURE1-8 Many-to-manyrelationshiptransformed
ThisisstandardoperatingprocedurewhenbuildinganERDandpreparingto
buildarelationaldatabase.
DatabaseNormalization
Gooddatabasedesignrequiresconsiderationfordatanormalization.Afull
analysisoftheconceptofdatabasenormalizationisbeyondthetaskofthisbook,
whichistoprepareyoufortheexam.Butitisworthnotingtherulesof
normalization,whicharerulesthatdrivethedesignofanysetoftables
composingarelationaldatabase.
Table1-4summarizesthemostcommonlevelsofnormalization.
TABLE1-4 LevelsofNormalization
Adatabaseadherestothefirstnormalform(1NF)whentablesarestructured
inaone-to-manyrelationship.Forexample,intheearlierexampleofshipsand
employees,itwouldhavebeenaviolationof1NFifyouhadinsteadplacedall
theshipnamesintheEMPLOYEEStableandrepeatedeachship’snamewith
eachrecordofeachemployeethatmighthappentobeassignedtothatship.By
separatingshipandemployeedata,yousatisfiedtherequirementfor1NF.
Secondnormalform(2NF)existswhennononkeyattribute(column)is
dependentuponpartofacompositekey.
Thirdnormalform(3NF)iswhentheone-to-manyrelationshipswithinadata
modelarestrictlyrespected—thatis,whenmany-to-manyrelationshipsareeach
transformedintoone-to-many-to-onerelationshipsasyoujustdidintheprevious
section.3NFrequiresthatprimarykeyvaluesareassociatedonlywithlogically
relatedattributes,andotherdata—suchaslookupdata—ismovedtoseparate
tables.
3NFdesignsareidealwherelotsofdataentryandupdatesareinvolved,as
theirdesignminimizesthepossibleduplicationofdataandtheassociatedriskof
dataconflicts,resultinginconflictingdatawithinthedatabase.Forexample,ifa
customeraddressisstoredinmultiplelocationsandthecustomercontactsthe
organizationwithanewaddress,youdon’twanttoruntheriskthatthenew
addressisrevisedinonelocationbutnotelsewhere.3NFreducesandeven
eliminatessuchrisk.
1NFand2NFfavorheavyread-onlyusecases.Forexample,ifyouneedto
producereportsthatconnect(orjoin)thecustomeraddresswiththecustomer
orderhistoryandtheitemsordered,thenthelessyourqueryhastogosearching
indifferenttables,thefasteryourreportwillexecute.Forsystemswithmillions
ofrecords,lotsofcomplexquerying,andnoneedtoaddnewdataorupdate
existingrecords,1NFand2NFofferfasterexecutionwithoutthedownsiderisk
ofconflictingdatathatmightexistina3NF.
Nothingintheexamobjectivesnorinmyexperienceofhavingtaken
theexamindicatesthatthetestinvolvesanythingotherthantypicalthird
normalformscenarios.
Forthesereasons,generallyspeaking,mosttransaction-baseddatabase
applicationsaredesignedtobe3NF.Datawarehousesandotherapplications
intendedtosupportanalysisandreportingaregenerallydesignedat2NFor1NF.
Many“bigdata”systemsaredesignedtoworkwith2NFor1NFmodels.
Mostdataentersanorganizationthroughatransaction-basedsystem(a3NF)
andisthenmovedintoadatawarehouseordatalakeforanalysis(a2NFor
1NF).Forthisreason,3NFisseenasastartingpointformostdata,and1NFor
2NFmodelsaresaidtobedenormalized,whichistosaythattheyadhereto
somelevelofnormalizationbelow3NF.“Bigdata”systemsareoftendescribed
asdenormalizedsincetheytendtoadhereto1NFor2NFdesign.
Thesedescriptionsaremerelyarefresherandarenotintendedtobean
exhaustiveanalysisofdatabasenormalization.Forthat,Ireferyoutoother
booksintheOraclePresslinethatdealwiththefundamentalsofdatabase
design.
Thereisnosinglerightorwrongwaytomodeleverysystem.Somedesign
decisionsinvolvetrade-offs,suchasbetterperformance(speedofresponse)
versuslessduplicatedata,oranefficientdatamodelversusamorecomplex
application.ThesearejustsomeofthechallengesadatamodelerorSQL
developerfaces.
CERTIFICATIONOBJECTIVE1.03
ExplaintheRelationshipBetweenaDatabaseand
SQL
Arelationaldatabaseisaphysicalrepresentationofabusinesssystem.Itis
sometimesdescribedasapersistentstore,meaningthatitisusedtoretain
informationthatmightbegeneratedorcapturedwhenasoftwareapplicationis
executing,butitretainsthatdataaftertheapplicationhasceasedexecution—the
data“persists”beyondthelifeofagivenapplication.Adatabasehousesdata
outsideoftheapplicationitselfandmakesthatdataavailabletoanyauthorized
inquiringsoftwareapplicationorservice.
SQListheindustry-standardlanguageforcreatingandinteractingwitha
relationaldatabase.SQLisamechanismtocreateandinteractwiththedatabase;
itisnotthedatabaseitself.SQLcommandscanbeusedonastand-alonebasis,
ortheycanbeinvokedfromwithinotherapplicationswritteninotherlanguages.
Mostpopularsoftwarelanguages,suchJava,PHP,andothers,havethe
capabilitytoissueSQLstatementstoarelationaldatabase.Thoselanguageswill
havetheabilitytotransfertheirowndataintoSQLstatements,sendtheSQL
statementtoadatabaseforprocessing,andthencapturethereturningdatafor
additionalprocessing.
Anumberoftoolsexistthatenableadeveloperorotherdatabaseusertoissue
SQLstatements.Forexample,Oracle’sSQL*PlusandSQLDevelopertools
empowerausertobuildandexecuteSQLstatementstobuildandinteractwith
databases.ThesetoolsissueSQLstatementsandreceiveoutputthesameasany
softwarelanguageorotherclientapplicationmightdo.
SohowdoyoupronounceSQL?Somepeoplepronounceitbyspellingout
theletters,asin“ess-cue-ell.”Mostofuspronounceitas“sequel.”Both
pronunciationsarefine,andbothareusedbyrespectedprofessionalsinthe
industry.Whateveryoudo,justdon’tcallit“squeal.”
Ioncehadaguycalltheofficelookingforajob,andheclaimedthathehad
fiveyearsofexperiencewith“Squeal.”Hedidn’tgetthejob.Don’tbethat
guy.
SQLisalanguageto
Createdatabasesandtheobjectswithinthem
Storedatainthosedatabases
Changeandanalyzethatdata
Getthatdatabackoutinreports,webpages,orvirtuallyanyother
useimaginable
TherearemanySQLstatements.Table1-5showssomeofthosethataremore
commonlyused.Foreachstatementshowninthistable,therearemanyclauses,
parameters,andotheradditionalfeaturesforeachone.Laterinthebook,you
willlookingreatdetailateachcommandthatiscoveredbytheexam.
TABLE1-5 SomeoftheMoreCommonlyUsedSQLStatements
AllSQLstatementsarecategorizedintooneofsixtypes.Table1-6showsthe
sixtypesofSQLstatementsinOracleSQL.Asyoucanseeinthetable,the
examignoresmanySQLstatements.Naturallyyouareconcernedonlywiththe
exam,sointhisbookyouwilllookatonlythreetypesofSQLstatements:Data
DefinitionLanguage,DataManipulationLanguage,andTransactionControl
Language.
TABLE1-6 SixTypesofSQLStatementsinOracleSQL
CERTIFICATIONOBJECTIVE1.04
DescribethePurposeofDDL
DataDefinitionLanguageconsistsofthoseSQLstatementsthatareusedto
builddatabaseobjects.Specifically,DDLstatementsareusedto
Create,alter,anddroptablesandotherdatabaseobjects
Addcommentsonaparticularobjecttobestoredinthedatabaseand
associatedwiththatobject
Issueprivilegestouserstoperformvarioustasksinthedatabase
Initiateperformanceanalysisonobjectsusingbuilt-intools
ThefollowinglistbrieflydescribesDDLstatementsthataretestedbythe
exam:
CREATEUsedtocreateauser,table,view,index,synonym,or
otherobjectinthedatabase.
ALTERUsedonanexistingobjectinthedatabasetomodifythat
object’sstructure,name,orsomeotherattribute.(Twoexceptionsarethe
usesofALTERwiththereservedwordsSESSIONandSYSTEM.
ALTERSESSIONandALTERSYSTEMarenottechnicallyconsidered
DDLstatementsbutfallunderadifferentcategory.Neitherisincludedon
thisexam.)
DROPUsedtoremoveadatabaseobjectfromthedatabasethathas
alreadybeencreatedwiththeCREATEstatement.
RENAMEChangesthenameofanexistingdatabaseobject.
TRUNCATERemovesalltherows—inotherwords,data—froman
existingtableinthedatabase.TRUNCATEissomethingofabrute-force
alternativetotheDELETEstatement,inthatTRUNCATEgivesup
recoveryoptionsofferedbyDELETEinexchangeforfaster
performance.ThesedifferencesinapproacharethereasonTRUNCATE
iscategorizedasDDLwhileDELETEisDML.
GRANTProvidesprivileges,orrights,touserobjectstoenable
themtoperformvarioustasksinthedatabase.
REVOKERemovesprivilegesthathavebeenissuedwiththe
GRANTstatement.
FLASHBACKRestoresanearlierversionofatableordatabase.
PURGEIrrevocablyremovesdatabaseobjectsfromtherecyclebin.
COMMENTAddscommentstothedatadictionaryforanexisting
databaseobject.
EachDDLstatementisrichwithoptionsandclauses.Iwillreviewand
provideexamplesofmostofthesestatementsasyouprogressthroughthebook.
SQLkeywords,suchasCREATE,arenotreally“commands”or“statements”
bythemselvesbutbecomecommandswhencombinedwithotherreserved
words,asinCREATETABLEorCREATESEQUENCE,whichare
commandsorstatements.Inpractice,CREATEcanbecalledastatementor
commandbyprofessionalsinthefield,andevenbyOracleCorporationin
variousformsofdocumentation.Technicallythereisadifference.However,
thisisn’tanissueontheexam.Similarly,thetermscommandandstatement
tendtobeusedinterchangeablybyOracle’sdocumentation.Ifyouconduct
somesearchesintheSQLLanguageReferenceManual,you’llfindplentyof
examplesofSQLstatementsbeingreferredtoascommands.Eitherisfine.
Andnoneoftheseissuesisaconcernontheexam.
CERTIFICATIONOBJECTIVE1.05
DescribethePurposeofDML
DMLreferstothosestatementsinSQLthatareusedtoworkwithdatainthe
objects.DMLstatementsareusedtoadd,modify,view,anddeletedataina
databaseobject,suchasatable.
ThefollowinglistbrieflydescribeseachDMLstatementthatistestedbythe
exam:
SELECTDisplaysdataofadatabasetableorview
INSERTAddsdatatoadatabasetable,eitherdirectlyor,insome
situations,throughaview
UPDATEModifiesexistingdatainatable,eitherdirectlyor,in
somesituations,throughaview
DELETERemovesexistingdatafromatable,eitherdirectlyor,in
somesituations,throughaview
MERGEPerformsacombinationofINSERT,UPDATE,and
DELETEstatementsinasinglestatement
TheSELECTstatementisratherinvolvedandwillgetseveralchapters’worth
ofreview.TheotherDMLstatementsarereviewedinthischapterandinvarious
sectionsthatfollow.
TransactionControlLanguage
InadditiontotheseDMLstatements,threemoreSQLstatementsareimportant
forworkingwithDML.ThesestatementsarenotpartofDMLbutinsteadare
categorizedasTCL.ThesestatementsarespecificallyidentifiedbyOracle
withinthecertificationobjectivesforDML,soI’lldiscusstheminthischapter.
Thesearethestatementsyouneedtostudy:
COMMITSavesasetofDMLmodificationsperformedinthe
currentdatabasesession
ROLLBACKUndoesasetofDMLmodificationsperformedduring
thecurrentdatabasesession
SAVEPOINTMarksapositioninasessiontoprepareforafuture
ROLLBACKtoenablethatROLLBACKtorestoredataataselected
pointinasessionotherthanthemostrecentcommitevent
InadditiontotheexplicitCOMMITcommand,anyDDLstatementwillcause
animpliedcommitactiontooccur.Inotherwords,ifyouissueanUPDATE
statementontableAandthenissueaGRANTstatementtouserBfor
accessingtableC,animpliedcommiteventwilloccur,andthedatayou
updatedtotableAwillbecommitted,eventhoughyouhaven’tissuedan
explicitCOMMIT.
CERTIFICATIONOBJECTIVE1.06
BuildaSELECTStatementtoRetrieveDatafroman
OracleDatabaseTable
Let’slookatasimpleexample.ConsidertheshipslistedearlierinTable1-2.A
validSQLcommandtocreateatableinwhichyoucouldstorethatinformation
mightlooklikethis:
Isay“might”looklikethisbecausethereareanumberofoptionsyoucould
includehere,includingprimaryorforeignkeydeclarations,datafiltering,
storageassignment,andotheroptionsthatgobeyondthissimpleexample.
You’lllookatmanyofthoseoptionslaterinthebook.Butthiscodesample
worksinanOracleSQLdatabase.
Next,here’saSQLcommandtoaddasamplerecordtothistable:
Again,thisisavalidcommand,albeitasimplifiedversion.Itinsertsone
recordofinformationaboutoneshipintothenewtableSHIPS.
Finally,let’screateaSQLcommandtodisplaythecontentsofthenewly
populatedSQLtable.
Ifallhasgonecorrectly,youshouldgetadisplaythatappearssomethinglike
thedisplayinFigure1-9.(Note:Figure1-9showstheoutputintheOracletool
SQLDeveloper.)
FIGURE1-9 OutputofthesampleSQLSELECTstatement
Asyoucansee,thedataisstoredinthetable,anditisstillthere.SELECT
merelydisplaysthedata;itdoesn’tchangethedataatall.Atitssimplestlevel,
thisiswhatSQLisallabout—writingstatementstocreatedatabaseobjectsand
thenworkingwiththoseobjectstostoreandretrievedata.
TheSELECTstatementisapowerfulSQLstatementthatcanbeusedto
querydatafromoneormoretablesinasinglestatement.AsingleSELECT
statementcanbeusedtotransformdata,aggregatedata,joinmultipletables,
filteroutunwanteddata,andmuchmore.Therestofthisbooklooksindetailat
avarietyofSQLstatementsastheypertaintothecertificationexam.
CERTIFICATIONSUMMARY
Thecertificationexamisatimed,multiple-choiceexamthatwilltestyouusing
entity-relationshipdiagrams,codeexamples,andmore.Itwillchallengeyou
withquestionspertainingtothecorrectuseofsyntax,thelogicalresultsofSQL
statements,andtheuseofvariousfunctionsandothercapabilitiesofSQL.This
bookistheidealtoolforexampreparation.Youshouldalsohaveaccesstoa
functionalinstallationofOraclesoftwarethatyoucanobtainfromOracle’sweb
site.
Theactofmodelingareal-worldbusinesssystemoftenincludesthecreation
ofadatamodel,whichisdepictedinanentity-relationshipdiagram.When
buildingarelationaldatabase,entitiesaretransformedintotables,andanentity’s
attributesareimplementedasthetable’scolumns.Agivenentity’srowsare
uniquelyidentifiedusingaprimarykey,andforeignkeyscanbeusedto
establisharelationshipbetweenonetableandanother.Databasenormalizationis
thepracticeofdesigningyourdatamodeltoserveyourbusinessapplicationat
anoptimallevel.Thethirdnormalformindicatesyourdatamodelisfreeof
duplicatedata,usesprimarykeystouniquelyidentifyasetofattributes,andhas
fullytransformedanymany-to-manyrelationshipsintoone-to-many
combinations.
SQListhemostwidelyusedlanguageforinteractingwithrelational
databases.Thelanguagecanbeusedtocreateandinteractwithdatabases.SQL
statementscanbeissuedfromwithinotherapplicationsandincorporatedwith
variouslanguages.
DDLstatementsarethoseSQLstatementsusedtocreatedatabaseobjects.
DMLstatementsareSQLstatementsusedtoworkwithexistingdatabase
objects.ThereareadditionalcategoriesofSQLstatements,includingTCL.
ASELECTstatementisthemostcommonlyusedSQLstatementandcanbe
usedtoquerydatafromoneormoretables,andsimultaneouslyitcantransform
andevenaggregatedatathatitreturns.
✓TWO-MINUTEDRILL
TheExam:AnOverview
Thischapterprovides
introductorymaterialthatisimportanttounderstandinpreparingforthe
exam.
The1Z0-071exam,OracleDatabaseSQLCertifiedAssociate,
whichisthesubjectofthisexamguide,has14certificationobjective
categories.Itshares11objectiveswithanotherexam,1Z0-051,SQL
FundamentalsI.But1Z0-071goesintomoredetailinthoseobjectives
andaddsthreemore.Also,1Z0-071tendstoemphasizetheobjectives
andsubobjectivesnotaddressedon1Z0-051.
Theexamincludes73questionsandallows100minutesto
completethem.Thatisanaverageofabout82secondsperquestion.
Thisbookwillprepareyoutostudyandsuccessfullytakeandpass
theexam.
Oracle’sSQLLanguageReferenceManualisoverkillasanexam
studyguide,asitcontainsfarmorethanyou’llneedfortheexam.Butit
isagoodreferencecompaniontothisbook.
IdentifytheConnectionBetweenanERDandaRelational
Database
TheERDisalogicalmodelofanexistingbusiness.
TheentitiesinanERDarethe“things”thatformthebasisofa
businessinformationsystemandthataretransformedintotablesina
database.
TheERDisalogicalrepresentationofabusiness;therelational
databaseisthephysicalmodelinwhichactualdatacanbehousedand
processedinsupportofthebusinessprocess.
ExplaintheRelationshipBetweenaDatabaseandSQL
Adatabaseisapersistentstoreofinformationthatcontinues—or
persists—beyondtheexecutionofagivenapplication.
SQListhemostwidelyusedlanguageforinteractingwitha
database.
SQLstatementscanbeissuedbyanyoneormoreapplicationstoa
singledatabase.Thoseapplicationscanbewritteninavarietyof
languagesthatsupportSQLcalls,includingsoftwarewrittenin
languagessuchasJava,PHP,andothers.
DescribethePurposeofDDL
DDLstatementsincludeCREATE,ALTER,andDROP.
DDLstatementsareasubsetofSQLandareusedtocreatenew
databaseobjectsoralterthestructureofexistingdatabaseobjects,
includingremovingthem.
DescribethePurposeofDML
DMLstatementsincludeSELECT,UPDATE,andINSERT.
DMLstatementsareasubsetofSQLandareusedtoworkwith
existingdatabaseobjects.
BuildaSELECTStatementtoRetrieveDatafromanOracle
DatabaseTable
TheSELECTstatementisthemostcommonlyusedSQLstatement.
SELECTcanbeusedtoquerydatafromasingletableorjoindata
togetherintwoormoretablestoreturndatathatislogicallyconnected.
SELECTcanbeusedwithfunctionsandothercapabilitiesto
transformdatainvariousways,whichisdiscussedatlengthlaterinthis
book.
SELFTEST
Thefollowingquestionswillhelpyoumeasureyourunderstandingofthe
materialpresentedinthischapter.Thefirstfivequestionstestyouonthe
backgroundinformationIdiscussedearliertoprepareyoufortheexam.The
remainingquestionsbegintotestyouonthespecificobjectivesoftheexam
itself.Furthermore,thesequestionsarewritteninthestyleandformatofthe
certificationexam,sotheycanbegoodpracticetohelpyougetgoing.Asisthe
casewiththeexam,someoftheseselftestquestionsmayhavemorethanone
correctanswer,soreadcarefully.Chooseallthecorrectanswersforeach
question.
TheExam:AnOverview
1.WhichofthefollowingtopicsarenotincludedintheSQL
FundamentalsIexambutareaddressedontheSQLAssociateexam?
(Chooseallthatapply.)
A.MERGE
B.Conversionfunctions
C.FLASHBACK
D.Externaltables
2.Ifyoufocusontryingtoachievetheminimumpassinggrade
requirementfortheexam,youcanstudymoreefficiently.
A.True
B.False
3.Theexamistimed.
A.True
B.False
4.The1Z0-071exam(whichisthesubjectofthisbook)hasbeen
officiallyvalidatedbyOracleCorporationagainstwhichofthefollowing
versionsoftheOracledatabase?(Chooseallthatapply.)
A.Everyversion
B.9i
C.11g
D.12c
5.Thebestexamguideyoucouldpossiblygetforpreparingtotakeand
passthe1Z0-071certificationexam,SQLAssociate,iswhichofthe
following?(Chooseallthatapply.)
A.Thisbook.
B.Thebookyouareholdingrightnow.
C.Thisherebook.
D.Don’tmakemetellyouagain.
IdentifytheConnectionBetweenanERDandaRelational
Database
6.WhentransforminganERDintoarelationaldatabase,youoftenuse
anentitytobuildadatabase’s:
A.Table
B.Column
C.Attribute
D.Relationship
7.Theuniqueidentifierofarowinadatabasetableisa(n):
A.ID
B.Primarykey
C.Primarycolumn
D.Column
ExplaintheRelationshipBetweenaDatabaseandSQL
8.WhichofthefollowingistrueofSQL?
A.Itisthemostcommonlyusedlanguageforinteractingwitha
database.
B.Itistheonlylanguageyoucanusetocreateadatabase.
C.Itistheonlylanguageyoucanusetointeractwithadatabase.
D.Noneoftheabove
9.WhatcanyouusetosubmitSQLstatementsforexecution?(Choose
allthatapply.)
A.PHP
B.Java
C.SQLDeveloper
D.SQL*Plus
DescribethePurposeofDDL
10.WhatisoneofthepurposesofDDL?(Choosethebestanswer.)
A.Querydatafromagiventable
B.Issueprivilegestousers
C.Removeexistingdatafromadatabasetable
D.Noneoftheabove
11.WhatcanDDLbeusedfor?(Choosethree.)
A.Addcommentstoadatabasetable
B.Addcolumnstoadatabasetable
C.Adddatatoadatabasetable
D.Addprivilegesforausertoadatabasetable
DescribethePurposeofDML
12.WhichoneofthefollowingisaDMLstatement?
A.ADD
B.ALTER
C.UPDATE
D.MODIFY
13.Whichofthefollowingcanbeusedtoremovedatafromatable?(Choose
two.)
A.DELETE
B.UPDATE
C.MODIFY
D.ALTER
BuildaSELECTStatementtoRetrieveDatafromanOracle
DatabaseTable
14.WhatcanaSELECTstatementbeusedtoquery?(Choosethebestanswer.)
A.Onlyonereport
B.Onlyonetable
C.Oneormorereports
D.Oneormoretables
15.WhichofthefollowingisnotacapabilityoftheSELECTstatement?
A.Itcantransformquerieddataanddisplaytheresults.
B.Itcanremovedatafromatable.
C.Itcanjoindatafrommultipletables.
D.Itcanaggregatedatabasedata.
SELFTESTANSWERS
TheExam:AnOverview
1. A,B,C,andD.SeeTable1-1forafulllistingofallthetopics
includedinbothoftheexams.
Alloftheanswersarecorrect.
2. B.Granted,itisasubjectiveissue,butOracleCorporation
specificallywarnsagainstthis.Onereasonisthatthepublishedminimum
requirementforapassingscorecanbechangedwithoutnotice.
Aisincorrect.Thestatementisfalse.
3. A.Yes,theexamistimed.
Bisincorrect.Theexamistimed,anditisimperativethatyoumanage
yourtimewellwhenyoutaketheexam.
4. CandD.Thetesthasbeenofficiallyvalidatedagainstthesetwo
versionsofthedatabaseandnottheothers.
AandBareincorrect.Theexamtestsforfunctionalitythatdidnotexist
inearlierversionsoftheOracledatabase.
5. A,B,C,andD.Duh.
Alloftheanswersarecorrect.
IdentifytheConnectionBetweenanERDandaRelational
Database
6. A.Thetableisaphysicalimplementationofanentity.
B,C,andDareincorrect.Acolumnisaphysicalimplementationofan
entity’sattribute.Arelationshipspecifieshowtwoentitiesrelatetoeach
other.
7. B.Theprimarykeyistheuniqueidentifierinatable.
A,C,andD.ThereisnoparticularunderstandinginERDsofrelational
databasesofanIDoraprimarycolumn.Bothmightbeusedtonamea
particularcolumnorotherobjectinthedatabase,buttheycarryno
particularspecificmeaning.Acolumnisanattributeofanentityina
relationaldatabase.
ExplaintheRelationshipBetweenaDatabaseandSQL
8. A.SQListhecommonlyusedlanguageforinteractingwith
databases.
B,C,andDareincorrect.SQLisnottheonlylanguageforcreatingor
interactingwithdatabases,butitisthemostcommonlyused.
9. A,B,C,andD.YoucanissueSQLstatementsfromanyofthelisted
options.
Alloftheanswersarecorrect.
DescribethePurposeofDDL
10. B.TheGRANTstatementispartofDDLandisusedtoissueprivilegesto
auser.
A,C,andDareincorrect.QueryingdataisperformedusingtheDML
statementSELECT.UsetheDMLstatementsUPDATEorDELETEto
removedatafromadatabasetable.UPDATEcanbeusedtoremovedata
fromoneormorecolumnsinagivenrow,andDELETEcanbeusedto
removeanentirerow.“Noneoftheabove”doesnotapplysinceoptionBis
correct.
11. A,B,andC.TheDDLstatementALTERcanbeusedtoaddcolumnsor
commentstoatable.UsetheDDLstatementGRANTtoaddprivileges.
Cisincorrect.ThereisnoDDLstatementtoremovedatafroma
databasetable.Forthis,useeitherUPDATEorDELETE.UPDATEcanbe
usedtoremovedatafromoneormorecolumnsinagivenrow,and
DELETEcanbeusedtoremoveanentirerow.
DescribethePurposeofDML
12. C.UPDATEisusedtochangedatainanexistingrow.
A,B,andDareincorrect.ThereisnoADDorMODIFYstatementin
SQL.ALTERisaDDLstatement.
13. AandB.TheDELETEstatementisusedtoremoveoneormorerows
fromagiventable.UPDATEisusedtochangedatainoneormoreexisting
rowsinagiventable,andthechangemayincludetheremovalofavaluein
acolumn,includingtheassignmentofNULLtoagiventable’scolumn
valueforthecandidateroworrows.
CandDareincorrect.ThereisnoMODIFYstatementinSQL.ALTER
isaDDLstatement.
BuildaSELECTStatementtoRetrieveDatafromanOracle
DatabaseTable
14. D.TheSELECTstatementcanbeusedtoqueryoneormoretables.
A,B,andCareincorrect.TheSELECTstatementdoesnotquery
reports;rather,itisusedtocreatereports.TheSELECTisnotlimitedto
queryingasingletable.
15. B.TheSELECTstatementisnotusedtoremovedatafromatable.Theact
ofqueryingmerelycopiesdataoutofatableforreportingandother
comparablepurposes,buttheSELECTstatement—asastand-alone
statement—alwaysleavesthedataitencountersuntouchedand
unmodified.However,itispossibletoincorporatefeaturesoftheSELECT
withinthecontextofotherstatements,suchasINSERTandUPDATE,
whicharedescribedlaterinthisbook.
A,C,andDareincorrect.TheSELECTstatementiscapableof
transformingdata,joiningdata,andaggregatingdata.
2
UsingDDLStatementstoCreateandManage
Tables
T
CERTIFICATIONOBJECTIVES
2.01CategorizetheMainDatabaseObjects
2.02CreateaSimpleTable
2.03ReviewtheTableStructure
2.04ListtheDataTypesThatAreAvailableforColumns
2.05ExplainHowConstraintsAreCreatedattheTimeofTableCreation
2.06DropColumnsandSetColumnsUNUSED
2.07CreateandUseExternalTables
✓Two-MinuteDrill
Q&ASelfTest
hischapterbeginstoexaminetables.Tablesaredatabaseobjectsthat
programmerscreatetostorethedatathatpopulatesthedatabase.
CERTIFICATIONOBJECTIVE2.01
CategorizetheMainDatabaseObjects
Databaseobjectsarethefoundationofanydatabaseapplication.Database
objectshouseandsupporteverythinganydatabaseapplicationneedsinorderto
formaworkingapplication.Thissectiontakesahigh-levellookatthedatabase
objectsthatcanbecreatedinanOraclerelationaldatabasemanagementsystem
(RDBMS),focusingonthoseobjectsthataretestedontheexam.Herewewill
separatedatabaseobjectsintocategoriesanddiscusstherelationshipbetween
objectsandschemas.Therestofthebookwilldelveintogreaterdetailoneach
databaseobjectthatisincludedintheexam.
WhatAreDatabaseObjects?
YoucancreatemanydifferenttypesofdatabaseobjectsintheOracleRDBMS.
Theexamlooksateightofthoseobjectsingreatdetail.
TheCompleteList
Adatabaseconsistsofoneormoredatabaseobjects.Thefollowinglistshows
objectsthatadatabasedevelopercancreateintheOracle12cRDBMS.Those
markedwithanasterisk(*)areincludedontheexam.(Source:SQLLanguage
ReferenceManual.)
Clusters Objecttables
Constraints* Objecttypes
Contexts Objectviews
Databaselinks Operators
Databasetriggers Packages
Dimensions Restorepoints
Directories Roles*
Editions Rollbacksegments
Externalprocedure
libraries Sequences*
Indexes* Stored
functions/procedures
Index-organizedtables Privatesynonyms*
Indextypes Publicsynonyms*
Javaclasses,etc. Tables*
Materializedviewlogs Tablespaces
Materializedviews Users*
Miningmodels Views*
Note:Nonschemaobjectsareitalicized;allothersareschemaobjects.Seethe
section“Schemas”formoreinformation.
Theexamdoesn’ttestforalloftheseobjects.Itignoresobjectssuchas
PL/SQLprogramunitsandJavaprogramunits,aswellasobjectsthatareof
moreinteresttodatabaseadministrators.Wewillconcernourselvesinthisbook
onlywiththosedatabaseobjectsthatareincludedintheexam.Thetypesof
databaseobjectsontheexamarelistedhere,inalphabeticalorder:
Constraints Synonyms
Indexes Tables
Roles Users
Sequences Views
ABriefDescription
Let’stakeabrieflookatthetypesofobjectsthatarethesubjectoftheexam:
TABLEAstructurethatcanstoredata.Alldataisstoredincolumns
androws.Eachcolumn’sdatatypeisexplicitlydefined.
INDEXAnobjectdesignedtosupportfastersearchesinatable.An
INDEXperformsmuchthesamewayasanindextoabook,bycopying
arelativelysmall,selectamountofinformation,sortingitforspeedy
reference,andtyingitbacktolocationsinthetableforsupportingquick
lookupsofrowsinthesourcetable.
VIEWA“filter”throughwhichyoucansearchatableandinteract
withatablebutthatstoresnodataitselfandsimplyservesasa“window”
ontooneormoretables.VIEWobjectscanbeusedtomaskportionsof
theunderlyingtablelogicforvariousreasons—perhapstosimplify
businesslogicortoaddalayerofsecuritybyhidingtherealsourceof
information.AVIEWcanbeusedtodisplaycertainpartsofatablewhile
hidingotherpartsofthesametable.
SEQUENCEAcounter,oftenusedtogenerateuniquenumbersas
identifiersfornewrowsastheyareaddedtoatable.
SYNONYMAnaliasforanotherobjectinthedatabase,oftenused
tospecifyanalternativenameforatableorview.
CONSTRAINTAsmallbitoflogicdefinedbyyoutoinstructa
particulartableabouthowitwillaccept,modify,orrejectincomingdata.
USERSThe“owners”ofdatabaseobjects.
ROLESAsetofoneormoreprivilegesthatcanbegrantedtoa
user.
Iwillrevieweachoftheseobjectsingreaterdetailthroughoutthebook.
Eachdatabaseobjectisconsideredtobeeithera“schemaobject”ora
“nonschemaobject.”Thisbegsthequestion,whatisaschema?
Schemas
Thissectiondescribesschemas—whattheyareandhowtheyrelatetodatabase
objects.
WhatIsaSchema?
Aschemaisacollectionofcertaindatabaseobjects,suchastables,indexes,and
views,allofwhichareownedbyauseraccount.Youcanthinkofaschemaas
beingthesamethingasauseraccount,butthereisaslightdifference—theuser
accounthousestheobjectsownedbyauser,andtheschemaisthatsetofobjects
housedtherein.Onedefinitionofschemathatyou’lloftenfindinOracle’s
documentation(andelsewhere)isthataschemaisa“logicalcollectionof
databaseobjects.”Technicallythat’strue,butitdependsonhowlogicaltheuser
choosestobewhenbuildingandplacingthoseobjectswithinhisorheruser
account.Ideallythereshouldbesomesensetowhyallthoseobjectsareinthere,
andideallya“schema”shouldn’tbejustarandomcollectionofobjects.
However,thereisnothingbuiltintotheOracleorSQLsystemsthatpreventsa
userfromdoingjustthat—randomlycollectingobjectsintoauseraccountand
thuscreatinga“schema”ofrandomobjects.Ideally,though,auseraccount
shouldbeseenandusedasalogicalcollectionofdatabaseobjects,drivenby
businessrulesandcollectedintooneorganizedentity—theschema.
Aschemahasthesamenameastheuseraccount.Auseraccountisidentified
byausernameandisassociatedwithasetofprivilegesandroles,whichare
coveredinChapter14,aswellasotherattributesnotaddressedontheexam.A
useraccount“owns”asetofzeroormoredatabaseobjects,andtogether,these
objectsconstituteaschema.Whileuseraccountsareoftenassociatedwitha
humanbeing,itisentirelypossibletocreateaschema(inotherwords,auser
account)whose“owner”isn’tahumanbeingatallbutperhapsisanapplication
process,someothersortofvirtualentity(perhapsaparticularbackground
process),orwhatevermakessensetosuitthebusinessrulesthatareinforce.So
inotherwords,oneuserwilloftenhaveoneuseraccountandthereforeone
schema.Buttheoppositeisn’tnecessarilytrue.Therecanbemoreuseraccounts
thanthereareactualusers.
Nowthatyouunderstandwhataschemaisandwhatauseraccountis,wecan
begintolookatdifferenttypesofdatabaseobjects,someofwhichareownedby
auserandarethereby“schema”objectsandsomeofwhicharenotschema
objectsbutarestilldatabaseobjectsnonetheless.
SchemaandNonschemaObjects
Alldatabaseobjectsfallintooneoftwocategories,ortypes.Thesetypes,asthe
Oracledocumentationcallsthem,areschemaandnonschema.
Table2-1showsthelistofbothschemaandnonschemaobjectsthatare
subjectsoftheexam.
TABLE2-1 Schemavs.NonschemaDatabaseObjects
Schemaobjectsarethoseobjectsthatcanbeownedbyauseraccount.
Nonschemaobjectscannotbeownedbyauseraccount.
Forexample,theUSERobjectisanonschemaobject.Auseraccountcannot
ownitself,anditcannotbeownedbyanotheruseraccount.Therefore,theUSER
object,whichisauseraccount,isanonschemaobjectandisapropertyofthe
databaseasawhole.ThesameistrueforROLEobjects.ROLEobjectsrepresent
oneormoreprivilegesthatcanbegrantedtooneormoreUSERobjects.Thus,a
ROLEinherentlyexistsataleveloutsideofanindividualUSERaccount—and
isthereforeanonschemaobject.APRIVATESYNONYMisownedbyauser
accountandisthereforeaschemaobject,butaPUBLICSYNONYMisa
variationontheSYNONYMobjectthatisownedbythespecialuseraccount
PUBLIC,whoseownedobjectsareautomaticallyavailabletotheentiredatabase
bydefinition.
AllotherobjectsareschemaobjectssuchasTABLE,INDEX,VIEW,andthe
otherslistedinTable2-1.
CERTIFICATIONOBJECTIVE2.02
CreateaSimpleTable
Theexamexpectsyoutobeabletorecognizethecorrectcodetocreateasimple
table.By“simple,”Oraclemeansthatyou’llberequiredtodefinethetable’s
name,columnnames,datatypes,andanyrelevantconstraints.
Tocreateatable,weusetheSQLcommandCREATETABLE.Theword
CREATEisaSQLreservedwordthatcanbecombinedwithjustaboutany
databaseobject(butnotall)toformaSQLcommand.Thesyntaxforthe
CREATEobjectTypestatementisshowninthiscodelisting:
where
objectTypeisareferencetoanobjectlistedinTable2-1.Hereare
someexceptions:
YoucannotcreateaCONSTRAINTinthisway.Also,whileCREATE
PRIVATESYNONYMobjectNameiscorrect,youusesimplyCREATE
SYNONYMobjectNameforpublicsynonyms,neverCREATEPUBLIC
SYNONYMobjectName,whichisincorrect.
objectNameisanameyouspecifyaccordingtothenamingrulesand
guidelinesdescribedlaterinthischapter.
attributesisanywherefromzerotoaseriesofclausesthatare
uniquetoeachindividualobjectType,whichwe’llreviewlater.
OneofthemostfrequentusagesoftheSQLcommandCREATEistocreatea
TABLE.Whenyoucreateatable,you’llalsocreatethetable’scolumnsand
optionallysomeassociatedobjects.
Let’slookatanexampleofabasicCREATETABLEstatement:
Ifyouweretoexecutethiscommandinaschemathatdidn’talreadyhaveatable
namedwork_schedule(that’simportant),thenyou’dgettheresultshownin
Figure2-1.
FIGURE2-1 ResultsofCREATETABLEwork_schedulestatement
Let’sanalyzethesyntaxoftheprecedingexampleofaCREATETABLE
statement:
ThereservedwordCREATE.
ThereservedwordTABLE.
Thenameofthetable,chosenbyyou,inaccordancewiththerules
ofnamingobjects,whichwereviewnext.
Apairofparentheses,inwhichareaseriesofcolumndeclarations,
eachseparatedbyacomma.Columndeclarationsconsistofthe
following:
Thenameofthecolumn,chosenbyyou,inaccordancewiththe
rulesofnamingobjects
Thedatatypeofthecolumn,takenfromthelistofavailabledata
types
Acommatoseparateeachcolumndefinitionfromthenext.
Asemicolontoendthestatement,asisthecasewithallSQL
statements.
Tofullyunderstandthesyntaxasjustdescribed,youneedtoexaminetwo
importantissues:therulesofnamingdatabaseobjectsandthelistofavailable
datatypes.Let’slookatnamingrulesnext;afterthatyou’lllearnaboutdata
types.
NamingaTableorOtherObject
Beforemovingonwiththedetailsofcreatingatable,let’stakealookatthe
rulesfornamingdatabaseobjects.Theserulesapplytotables,views,indexes,
andalldatabaseobjects—includingatable’sconstraints,ifanyarecreated.The
samenamingrulesalsoapplytoatable’scolumns.
Alltableshaveaname.Eachtableconsistsofoneormorecolumns,andeach
columnhasaname.(Forthatmatter,eachdatabaseobjectinthedatabasehasits
ownname—eachindex,view,constraint,synonym,andobjectinthedatabase
hasaname.)
WhenyouusetheSQLkeywordCREATEtocreateadatabaseobject,you
mustcomeupwithanameandassignittotheobject,andsometimes—asinthe
caseofatable—toindividualcomponentswithintheobject,suchasthecolumns
ofatable.
Thereisaspecificsetofrulesfornamingdatabaseobjects—includingtables,
tablecolumns,views—anythingyoumustnameinthedatabase.Thenext
sectionlooksatthosenamingrules.
NamingRules—Basics
Therulesfornamingtables,andanydatabaseobject,includethefollowing:
Thelengthofthenamemustbeatleast1characterandnomorethan
30characters.
Thefirstcharacterinanamemustbealetter.
Afterthefirstletter,namesmayincludeletters,numbers,thedollar
sign($),theunderscore(_),andthepoundsign(#),alsoknownasthe
hashmarkorhashsymbol.Nootherspecialcharactersareallowed
anywhereinthename.
Namesarenotcase-sensitive.
NamescannotbereservedwordsthataresetasideforuseinSQL
statements,suchasthereservedwordsSELECT,CREATE,andsoon.
SeethefollowingcompletelistofreservedwordsfromOracle’sSQL
LanguageReferenceManual.Thesewordsareoff-limitswhenyoucreate
namesforyourdatabaseobjects.
Theserulesareabsolute.Ifyouattempttocreateatableoranyotherdatabase
objectwithanamethatviolatestheserules,theattemptwillfail,you’llreceive
anerrorcodefromthedatabase,andyourobjectwillnotexist.
QuotedNames
WhilenotrecommendedbyOracle,itispossibletocreatedatabaseobjectnames
thatare“quoted.”Aquotednameissurroundedwithdoublequotes,suchasthe
"Companies"object.Here’sanexample:
Quotednamesmustalwaysbereferencedwiththeirquotationmarks.
Quotednamesvaryfromnonquotedinseveralways.Forexample,quoted
namesmaybeginwithanycharacter.Namesmayincludespaces.
Quotednamesarecase-sensitive.Forexample,nowthatI’vecreatedatable
called"CompanyEmployees",thisquerywillwork:
Butthefollowingwillnot:
Quotednamesmayincludereservedwords.Andtoreiterate,quotednames
mustalwaysbereferencedwiththeirquotationmarks.
Generally,mostdatabaseobjectsarecreatedwithouttheuseofdouble
quotationmarks,andOracleadvisesagainsttheiruse.Butyoumayneedto
knowaboutthemanyway.
UniqueNamesandNamespaces
Whathappensifyoutrytocreateadatabaseobjectwithanamethatmatchesthe
nameofanotherdatabaseobjectthat’salreadyinthedatabase?Canyoudoit?
Whathappenstotheexistingdatabaseobject?Willyoubeabletousethe
resultingdatabaseobject?Theansweristhatitdependsonyourobject’s
relationshiptotheotherobjectthatalreadyexists,aswellasonsomethingcalled
thenamespace.
Anamespaceisalogicalboundarywithinthedatabasethatencompassesa
particularsetofdatabaseobjects.Thereareseveralnamespacesatworkatany
giventime,dependingonthecontextinwhichyouareworking.
Understandingthenamespaceisnecessaryinordertounderstandwhether
youmayormaynotspecifyduplicatenamesforanyparticulardatabaseobject.
SeeFigure2-2foradiagramthatdemonstratesthenamespaceboundaries.Note
thateachsquareenclosesadifferentnamespace.InFigure2-2,thereareseveral
namespacesidentified:
FIGURE2-2 Diagramofnamespaceboundaries
USERandROLEobjectsareintheirowncollectivenamespace.
PUBLICSYNONYMobjectsareintheirownnamespace.
TABLE,VIEW,SEQUENCE,PRIVATESYNONYM,anduser-
definedTYPEobjectshavetheirowncollectiveuniquenamespace
withinagivenschema.(NotethatIhaven’tpreviouslymentioneduser-
definedtypes.Theyaren’tontheexam.Iincludetheminthisdiscussion
tobecomplete,butyouwon’tneedtounderstandthemfortheexam.)
INDEXobjectshavetheirownnamespacewithinagivenschema.
CONSTRAINTobjectshavetheirownnamespacewithinagiven
schema.
Whatallofthismeansisthatyoumustprovideuniquenamesforanobject
withinitsownnamespace.Objectsthatshareanamespacemusthaveunique
nameswithinthatnamespace.Objectsindifferentnamespacesareallowedto
haveidenticalnames.Herearesomeexamples:
IfyoucreateatableinoneschemacalledWORK_SCHEDULE,
thenyoucannotcreateanothertablecalledWORK_SCHEDULEwithin
thatsameschema.Butyoucandoitinanotherschema,providedthere
isn’talreadyaWORK_SCHEDULEinthatschema.
Let’ssayyouhaveaschemacalledHR(asinHumanResources).In
theHRschema,youcancreateatablecalled,say,PAYROLL.You
cannotcreateaVIEWinthatsameschemacalledPAYROLL.Butyou
cancreateanINDEXcalledPAYROLL.Youcanalsocreatea
CONSTRAINTcalledPAYROLL.ButyoucannotcreateaSEQUENCE
calledPAYROLL.
Intheentiredatabase,eachUSERobjectmusthaveauniquename
andsomusteachROLEobject.
Notethatlateryou’llseeyoucancreateaTABLEandgiveitaprimarykey
CONSTRAINT.Ifyoudothis,you’llhavetheoptionofnamingthat
CONSTRAINT.Ifyoudo,thesystemwillautomaticallycreateanINDEXfor
theCONSTRAINT,anditwillnametheINDEXwiththesamenameasthe
CONSTRAINT.Youcanoverridethisandassignyourownnamewiththe
USINGINDEXclauseoftheCREATETABLEstatement,somethingyou’ll
learnaboutlater.
Whennamingobjects,choosedescriptivenamesthatcanbepronounced.Be
consistent:ifyourtablesofEMPLOYEES,CUSTOMERS,andVENDORS
eachincludeareferencetoaperson’sname,makethosecolumnnamesallthe
same—NAME,LAST_NAME,andFIRST_NAME,whatever.Justbe
consistent.Considernamingalltablesinpluralform,sinceeachmostlikely
containsmultiplerecords.Considerusingastandardprefixforeverydatabase
objectthat’sassociatedwithaparticularapplication.Forexample,fora
HumanResourcesapplication,prefixeachtablewithHR_,butavoidusing
prefixesthatOracleCorporationusesforitssystem-definedobjects:SYS_,
ALL_,DBA_,GV$,NLS_,ROLE_,USER_,andV$.
System-AssignedNames
You’llseeabitlaterthatyoumaycreateanobjectindirectly.Thishappens,for
example,whenyoucreateatableandwithintheCREATETABLEstatement
youoptionallydefineanassociatedconstraint,butwithoutprovidinganamefor
theCONSTRAINT.ThelanguagesyntaxoftheCREATETABLEstatement
allowsthistohappen,asyou’llsoonsee,andtheresultisnotonlyyournewly
created—andnamed—table,butalsoanewlycreatedconstraintforthattable.
Somedevelopersrefertotheseconstraintsasanonymous,buttheyaren’t
anonymousatall.Thesystemwillautomaticallygenerateanameforthat
constraint—anamelikeSYS_C001234,withaprefixofSYS_C,followedbya
numbergeneratedautomaticallybythesystem.
TheSQLStatementCREATETABLE
TheSQLstatementCREATETABLEisacomplexstatementwithmanyclauses
andparameters.Theexamtestsforonlysomeofitsfunctionality,includinghow
tocreatecolumns,specifydatatypesforthosecolumns,andcreateconstraints.
Here’sanexampleofarelativelysimpleCREATETABLEstatement:
Inthisexample,youarecreatingatablewithsevencolumnsandtwoconstraints.
Eachofthecolumnsisgivenanameandadatatype.Thedatatypesprovide
somerulesandrequirementsforthedatathat’senteredintothecolumns.For
example,onlynumberscanbeenteredintoCRUISE_TYPE_ID.Onlydate
valuescanbeenteredintoSTART_DATE.
NotethattheSTATUScolumnhasadefaultvalueof'DOCK'.IfanINSERT
statementomitsthevalueforSTATUS,then'DOCK'willbethevaluefor
STATUS.
AttheendoftheCREATETABLEstatementisanadditionallinethatcreates
aCONSTRAINT.ThisparticularCONSTRAINTdefinestheCRUISE_ID
columnasaprimarykey,whichmeansthatanyrowaddedtotheCRUISES
tablemustincludeavalueforCRUISE_ID,andthatvaluemustbeunique—it
cannotduplicateanypreexistingvalueforCRUISE_IDinanyotherrowalready
presentinthetable.
There’salsoaNOTNULLconstraintthat’sappliedtotheCAPTAIN_ID
column.ThatCONSTRAINTisn’texplicitlynamed,butit’saCONSTRAINT
nonetheless,anditwillbeassignedasystem-generatedname.
Intheremainderofthischapter,you’lllearnhowtoreviewthestructureofa
table.You’lllookatthedifferentdatatypesyoucanusetocreatecolumnsina
table.Thechapterwillconcludebylookingatconstraintsandhowtheycanbe
createdatthetimeyoucreateatable.
CERTIFICATIONOBJECTIVE2.03
ReviewtheTableStructure
Onceyouhavecreatedatablesuccessfullyinthedatabase,youcanreviewthe
table’sstructurewiththeDESCRIBEstatement.TheDESCRIBEstatement,
oftenabbreviatedasDESC,isn’taSQLstatement;it’saSQL*Plusstatement
thatisuniquetoOracle.(Someotherproductvendorshavesinceimplemented
DESCintheirownSQLproducts.)TheDESCstatementisn’taSQLstatement
—itisnotdescribedintheOracleDatabaseSQLLanguageReferencemanual.
DESCfallsunderthecategoryofoneoftheseveralSQL*Plusenhancements,
uniquetoOracle,anddocumentedinOracle’sSQL*PlusUser’sGuideand
Referencemanual,whichdetailstheDESCstatement.Buteventhoughitisnot
consideredSQL,DESCisimportanttounderstandsinceDESCisusefulfor
quicklyreviewingatable’sstructure.
OnenoteaboutSQL*Plusstatements:theydon’trequireasemicolonatthe
end.You’llseeDESCstatementswithoutasemicolonintheexamplesthat
follow.NotethataSQL*Plusstatementconcludesattheendofitsfirstline
unlessacontinuationcharacterisplacedattheendofthelinetoexplicitly
indicatethatthestatementcontinuestothenextline.Thisisasignificant
differencefromSQLstatements,whichrequireasemicolonattheirend,and
continuetomultiplelinesuntilthesemicolonisencountered.TheSQL*Plususe
ofthecontinuationcharacterandthefullsyntaxofSQL*Plusstatementsare
beyondthescopeofthisbook.TheonlypointtonotehereisthatSQL
statementscontinueuntilasemicolonmarkstheend,spanningmultiplelinesif
required.ButaSQL*Plusstatementdoesnotbehavethatwayandendsonits
firstline,withorwithoutasemicolon,unlessacontinuationcharacterisused,
andwewon’tseetheneedforthatinourexamplesthatfollow—we’lljustsee
theDESCstatementendwithoutasemicolon.Alsonote:youcanactuallyplace
asemicolonattheendofaSQL*Plusstatementand—technically—itwillhave
noeffect.Youwon’tgetanerrormessage,norisitrequired.Butdon’tforget:a
semicolonisalwaysrequiredattheendofaSQLstatement.
Let’stakealookatanexampleofDESC.ConsidertheCREATETABLE
CRUISESstatementyousawearlier:
AssumingthisSQLstatementwasexecutedinthedatabasesuccessfully,
resultinginthetableCRUISESbeingstoredinthedatabase,thenyoucould
issuethefollowingSQL*Pluscommand:
DESCcruises
Figure2-3displaystheresult.Noticetheoutputlistshowsathree-column
display.
FIGURE2-3 ResultofthecommandDESCcruises
Thefirstcolumnintheoutputlistingistitled“Name”andshowsthe
table’scolumnnamesthatyouspecifiedwiththeCREATETABLE
statement.
Thesecondcolumnintheoutputlistingistitled“Null”andshows
whetherthereisaNOTNULLconstraintappliedtothatparticular
columninthetable.Inotherwords,willanyrowthat’saddedtothe
databasebeallowedtoomitthisparticularvalue?
Thethirdcolumnintheoutputlistingistitled“Type”andshowsthe
datatypefortheparticulartable’scolumninquestion.
Forexample,theDESCCRUISESoutputshowsthattheCRUISEStablehas
acolumntitledCAPTAIN_ID,itsdatatypeisNUMBER,andithasaNOT
NULLCONSTRAINTappliedtoit.
CERTIFICATIONOBJECTIVE2.04
ListtheDataTypesThatAreAvailableforColumns
ThissectionlistsandexplainsdatatypesprovidedbyOraclethatcanbe
assignedtocolumnsinatable.We’lllookatexamplesinlaterchapters—for
nowwe’reinterestedonlyinlistinganddescribingthem.
DatatypesareassignedtodifferenttypesofobjectsintheSQLdatabaseand
throughouttheOraclesystem.Inatable,eachcolumnmustbeassignedadata
type.Acolumn’sdatatypedefineswhatsortofinformationis—andisnot—
acceptedasinputintothecolumn.Itdetermineshowthevaluesinthecolumn
canbeused,howtheywillbehavewhencomparedtoothervaluesorevaluated
inexpressions,andhowtheyaresorted.
Oracle’sowndocumentationreferstodatatypesasbothdatatypesand
datatypes.Thesetwoexpressionsarethesamething.
Mostdatatypesfallunderoneofthegeneralcategoriesofnumeric,character,
ordate.There’smoretoitthanthis,butmostdatatypesfallintooneofthese
threegeneralcategories.Inadditiontothesethreeisacategoryreferredtoas
largeobject,orLOB,datatypes.LOBscanincludecharacterdatabutcannotbe
includedinaprimarykey,DISTINCT,GROUPBY,ORDERBY,orjoins.
Character
Characterdatatypesarealsoknownastextorstringdatatypes,andtheyinclude
thefollowing:
CHAR(n)ThenameCHARisshortfor“character.”Thisisafixed-
lengthalphanumericvalue.Anyalphanumericcharacterisacceptedas
input.Thenindicateshowlongthevaluewillbe.TheCHAR(n)data
typeacceptsvalidinputandpadsanyremainingunusedspacewith
blankstoensurethatthelengthofyourvaluewillalwaysequalthevalue
ofn.Forexample,ifyoudeclareacolumnwithdatatypeofCHAR(5),
thenavalueof,forexample,“A”willbestored—andretrieved—as“A”
whereAisfollowedbyfourblankspaces.Anyattempttoenteravalue
thatislongerthannwillresultinanerror,andthevaluewillnotbe
accepted.WhendeclaringaCHARdatatype,theinclusionofnis
optional;ifitisomittedinthedeclaration,avalueof1isassumed.For
example,astatementofCREATETABLEcruises(cruise_idCHAR)will
resultinaCRUISE_IDcolumnwithadatatypeofCHAR(1).The
maximumallowedvaluefornis2000.
VARCHAR2(n)ThenameVARCHARissortofanabbreviationfor
“variablecharacter.”Thisisavariable-lengthalphanumericvalue.Then
indicatesthemaximumallowablelengthofthevaluestoredwithin,but
contrarytoCHAR,theVARCHAR2formatwillnotpaditsvalueswith
blanks.Itslengthvariesaccordingtothedataitcontains.Alsodifferent
fromtheCHARdatatypeisthatVARCHAR2requiresntobespecified.
Theminimumvalueofnis1;themaximumallowablelengthof
VARCHAR2is4000.(Note:Theissueofamaximumvaluein
VARCHAR2isactuallyabitmorecomplexthanthis—themaximumis
technically4,000bytesandnotreally4,000characters,andbydefault
mostOracledatabaseimplementationsareconfiguredsothatone
characterequalsonebyte.Butit’spossibletooverridethis,whichwould
theoreticallychangethemaximumnumberyoucanusefornina
VARCHAR2declaration.Furthermore,beginningwithOracle12c,there
isaninitializationparameterMAX_STRING_SIZEthatcanbesettothe
valueEXTENDED,enablingVARCHAR2declarationsofupto32,767.
Forourpurposeshereitdoesn’treallymatter—ithasn’tbeenanissueon
theexam.Butkeepaneyeoutforthis;itmaybeintroducedintothe
examinthefuture.)
Numeric
Numericdatatypesincludethefollowing:
NUMBER(n,m)Thisdatatypeacceptsnumericdata,including
zero,negative,andpositivenumbers,wherenspecifiesthe“precision,”
whichisthemaximumnumberofsignificantdigits(oneithersideofthe
decimalpoint),andmisthe“scale,”meaningthetotalnumberofdigits
totherightofthedecimalpoint.Bothnandmareoptional;ifbothare
omitted,bothdefaulttotheirmaximumvalues.Ifnisspecifiedandmis
omitted,mdefaultstozero.Thevalueforncanrangefrom1to38;the
valueformcanrangefrom–84to127.Notethatthesearenotthelargest
valuesyoucanhavebutratherthelargest(andsmallest)specifications
forvaluesyoucanhave—Oracle’sSQLLanguageReferenceManual
carefullystatesthatthevaluesacceptedforaNUMBERdatatyperange
from1.0×10–130upto“butnotincluding”1.0×10126.Ifavalueentered
intoaNUMBERcolumnhasaprecisiongreaterthanthespecifiedvalue,
anerrormessagewillresult,andthevaluewillberejected.Ontheother
hand,ifavalueisenteredthatexceedsthedeclaredscale,theentered
valuewillberoundedoff(.5isroundedup)andaccepted.Also,a
negativevalueformidentifieshowmanysignificantdigitstotheleftof
thedecimalpointwillberoundedoff.SeeTable2-2foranexampleof
howallofthisworks.It’sconsideredgoodpracticetospecifythe
precisionandscaleaspartoftheoveralldataintegritychecktoplace
someboundariesaroundthelimitsofwhatthebusinesslogicoftheintent
ofthecolumnwillaccept.
TABLE2-2 ExamplesofNUMBERPrecisionandScale
Date
DatedatatypesaresometimesreferredtoinOracle’sdocumentationas
datetimes.Eachdatedatatypeconsistsoffields,andeachfieldisacomponent
ofadateortime,suchashours,minutes,orthemonthvalue,andsoon.See
Table2-3foralistofthefieldsthatareusedinvariouscombinationstoform
datedatatypes.
TABLE2-3 DatetimeFields
Thedatatypesthatsupportdateandtimeinformationincludethefollowing:
DATEThisacceptsdateandtimeinformation.Thefieldsstored
includeyear,month,date,hour,minute,andsecond.Datevaluesmaybe
storedasliteralsorusingconversionfunctionsthatyou’llseelaterin
Chapter6.Dateliteralsareenclosedinsinglequotationmarksandmay
bespecifiedinanumberofways.ThedefaultOracledateformatfora
givencalendardayisdefinedbytheparameterNLS_DATE_FORMAT.
TheinitializationparameterNLS_DATE_FORMATspecifiesthe
defaultdateformatforyourdatabase.However,itisalsopossiblethat
NLS_DATE_FORMATisnotexplicitlysetbutisinsteaddefaultedtothe
associateddateformatofanotherinitializationparameter,
NLS_TERRITORY.Toseethevaluespecifiedfortheinitialization
parameterNLS_TERRITORYinyourdatabase,useSHOW
PARAMETERSNLS_TERRITORY.Toseewhetheravaluefor
NLS_DATE_FORMATwasspecifiedforyourdatabase,useSHOW
PARAMETERSNLS_DATE_FORMAT.Regardlessofwhetherthe
NLS_DATE_FORMATwasexplicitlysetforyourdatabase,youcansee
itsvaluewiththisquery:
TheNLS_DATE_FORMATparametercanbechangedwithALTER
SESSIONorALTERSYSTEM,whichareSQLstatementsthatarenot
includedontheexam.Bydefault,installationsintheUnitedStatesand
UnitedKingdomusetheNLS_DATE_FORMATofDD-MON-RR,
whereDDisthetwo-digitday,MONisthethree-letterabbreviationfor
themonth,andRRisthetwo-digityear,wherevaluesofRRranging
from00to49areassumedtobeinthetwenty-firstcentury(2000to
2049),whileRRvaluesrangingfrom50to99areassumedtobeinthe
twentiethcentury(1950through1999).Forexample,'01-NOV-10'isthe
firstofNovemberin2010.(Note:ThesamedateinANSIformatis
'2010-11-01'.ANSIformatis'YYYY-MM-DD',whereYYYYisthe
four-digityear,MMisthetwo-digitmonth,andDDisthetwo-digitday.)
SeeChapter6formoreonthesetopics.
TIMESTAMP(n)ThisisanextensionofDATEthataddsfractional
secondprecision.TIMESTAMPstoresyear,month,day,hours,minutes,
seconds,andfractionalseconds.Thevaluefornspecifiestheprecision
forfractionalseconds.Therangefornis0–9.Ifnisomitted,itdefaults
toavalueof6.
TIMESTAMP(n)WITHTIMEZONEThisisavariationof
TIMESTAMPthataddseitheratimezoneregionnameoranoffsetfor
timezone.TIMESTAMPWITHTIMEZONEisusedintrackingdate
informationacrossdifferenttimezonesandgeographicalareas.The
rangefornis0–9.Ifnisomitted,itdefaultstoavalueof6.
TIMESTAMP(n)WITHLOCALTIMEZONEThisisavariation
ofTIMESTAMP.TheTIMESTAMPWITHLOCALTIMEZONEdiffers
fromTIMESTAMPWITHTIMEZONEinthatthetimezoneoffsetis
notstoredwiththecolumn’svalue,andthevalueretrievedissenttothe
userintheuser’slocalsessiontimezone.Inotherwords,whiletheoffset
isnotstoredwithTIMESTAMPWITHLOCALTIMEZONE,the
systemcalculatestheoffsetautomaticallywhenthevalueisaccessed
fromadifferenttimezone.Endusersseethetimeintheirownlocaltime
zone,regardlessofthetimezonewherethedatabaseserverhappensto
reside.Theoffsetiscalculatedautomatically.Ifnisomitted,itdefaultsto
avalueof6.
INTERVALYEAR(n)TOMONTHThisstoresaspanoftime
definedinonlyyearandmonthvalues,wherenisthenumberofdigits
usedtodefinetheYEARvalue.Therangeofacceptablevaluesfornis
0–9;thedefaultfornis2.Thisdatatypeisusefulforstoringthe
differencebetweentwodatevalues.
INTERVALDAY(n1)TOSECOND(n2)Thisstoresaspanoftime
definedindays,hours,minutes,andseconds,wheren1istheprecision
fordays,andn2istheprecisionforseconds.Therangeofvaluesforn1
is0–9,andthedefaultis2.Thevalueforn1specifieshowmanydigits
areacceptedindeclaringthesizeofanumberforDAYtobespecified.
Thevalueforn2isthefractionalsecondsprecisionforSECOND;the
rangeforacceptablevaluesis0–9,andthedefaultis6.Thisisusefulfor
storingthedifferencebetweentwodatevalues.
Thedatatypesthathandletimezonedifferencesareimportanttothe
exam.You’lllookatfunctionsinChapter6thatdealwiththesedata
types.
LargeObjects
Theexamwilladdressyourknowledgeofhowtouselargeobject,orLOB,data
typesinSQLlanguagesyntax.LOBscangenerallybeusedlikeotherdatatypes.
TablesmayhavemultiplecolumnswithLOBdatatypes.However,LOBscannot
beprimarykeys,andtheycannotbeusedwithDISTINCT,GROUPBY,
ORDERBY,orjoins.
LOBsincludethefollowing:
BLOBThenameBLOBisanabbreviationfor“binarylargeobject.”
BLOBacceptslargebinaryobjects,suchasimageorvideofiles.
Declarationismadewithoutprecisionorscale.(Themaximumsizeis
calculatedbywayofaformulathatincludesseveralitemsnotonthe
exam,includingastartingsizeof4GB,somethingcalledtheCHUNK
parameter,andthesettingforthedatabaseblocksize,whichisasetting
thataffectsallstorageinthedatabase.)
CLOBThenameCLOBisanabbreviationfor“characterlarge
object.”CLOBacceptslargetextdataelements.Declarationismade
withoutprecisionorscale,andthemaximumsizeiscalculatedinthe
samemannerthatitisfortheBLOBdatatype.
NCLOBThisacceptsCLOBdatainUnicode.Themaximumsizeis
calculatedinthesamemannerthatitisfortheBLOBdatatype.
RegardingUnicode—itisacharactersetthatservesasanalternativeto
ASCIIandrepresentsamoreuniversalstandardthatsupportsallmajor
languagesmoreeasilythantheotherimplementationsinusetoday.
OracleandmostothermajorvendorshaveadoptedUnicodeintotheir
products,andcommonwebtechnologiesalreadysupportit.Giventhe
increasingroleofglobalizationandmultilanguagesupport,anylegacy
applicationdeployedwithoutUnicodemayinevitablyrequirea
conversioneffortdowntheroad.OracleCorporationisofficially
recommendingtheuseofUnicodeasthedatabasenationalcharacterset
forallnewsystemdevelopment.
Here’sanexampleofatablethatincludesacolumnoftheCLOBdatatype:
Theprecedingexamplecreatesatablewithtwocolumns,thesecondofwhich
isaCLOB.Thatcolumncanreceiveextremelylargetextdataasinput.
OracleCorporationdiscouragestheuseoftheoldLONGdatatypeand
encouragesyoutoconvertLONGdatatypestoLOBdatatypes,whichhave
fewerrestrictions.Forexample,youcanaddmorethanoneLOBcolumntoa
table,youcanselectLOBcolumns,youcaninsertintotableswithLOB
columns,andyoucandeleterowswithLOBvalues.However,asIstatedearlier,
youcannotuseLOBsinGROUPBYorORDERBYclauses.
Allofthedatatypesyou’veseensofararebuiltinbyOracleandincludedwith
SQL.Allofthesedatatypesareknownas“built-in”datatypes.However,it’s
possibleforuserstocreatetheirownunique“user-defined”datatypes.User-
defineddatatypesarecreatedusingtheSQLstatementCREATETYPE.They
areusedinPL/SQLcodeandarenotasubjectoftheexam.
CERTIFICATIONOBJECTIVE2.05
ExplainHowConstraintsAreCreatedattheTimeof
TableCreation
YoucancreateaCONSTRAINTtosupportotherobjects,specificallyTABLE
objects.Butthereisn’taCREATECONSTRAINTstatement;youcreatea
CONSTRAINTaspartofanotherstatement,suchasCREATETABLEor
ALTERTABLE.WhenyoucreateaCONSTRAINT,youcanchoosetonameit
yourselforletthesystemautomaticallyassignaname.Thecreationofcertain
constraints—PRIMARYKEY,UNIQUE,andFOREIGNKEY—will
automaticallytriggerthecreationofacorrespondingindexobjectofthesame
name.
Here’sanexampleofaCREATETABLEstatementthatincludesthesyntax
tocreateaCONSTRAINT:
Intheprecedingexample,wecreateaTABLEcalledPOSITIONS,which
consistsofthreecolumns,POSITION_ID,POSITION,andEXEMPT.Afterthe
EXEMPTcolumnisdefined,thisparticularexampleshowsanadditionallineof
codetocreateaCONSTRAINT.(Thereareotherformatsforcreating
constraints,asyouwillsoonsee.)Inthisexample,wearechoosingtonamethe
CONSTRAINT,somethingthatwedon’tnecessarilyhavetodo.Thenamewe
areassigningtothisCONSTRAINTisPOSITIONS_PK.We’respecifyingthat
thisCONSTRAINTisoftypePRIMARYKEY,andwe’reapplyingthe
CONSTRAINTtothecolumninthistablethat’scalledPOSITION_ID,which
wedefinedfirst.
Let’slookatsomespecificsnext.
CreatingCONSTRAINTSintheCREATETABLE
Statement
TherearetwowaysinwhichaCONSTRAINTcanbecreatedatthetimeof
TABLEcreation:inlineandoutofline.
CREATETABLE:In-LineConstraints
HereisanexampleofhowtocreateaPRIMARYKEYconstraintinline:
Inthisexample,wecreateaPRIMARYKEYconstraintonthecolumn
PORT_ID.TheconstraintonPORT_IDisspecifiedaspartofthePORT_ID
declaration.Itis“inline”withthecolumnuponwhichtheconstraintisapplied.
WhenIsay“inline,”Imeanthatafterwedeclareacolumntowhichwewantto
applyaconstraint,weincludetheconstraintspecification(inthiscase,fora
primarykey)beforethecommathatseparatesthecolumnspecification
(PORT_IDinthisexample)fromthenextcolumnspecification,whichis
PORT_NAME.Thefactthattheexamplehappenstoincludethewords
PRIMARYKEYonthesamelineasPORT_IDisincidentalandunnecessaryto
makethisan“inline”format.
Inthisexample,weomittheinclusionofaspecifiednamefortheconstraint,
whichcausestheOracledatabasetogenerateaconstraintnameautomatically,of
theformSYS_Cn.Here’sanexample:
You’llseehowyou’llbeabletoidentifythesystem-assignednamewhenyou
lookatthedatadictionaryinChapter12.
Alternatively,youcanoptionallyspecifytheconstraintnameatthetimeof
creation,byprecedingthereservedwordsPRIMARYKEYwiththereserved
wordCONSTRAINT,followedbyanameyoumakeupaccordingtotherulesof
namingdatabaseobjects,likethis:
Thesetwoapproachesarereferredtoasin-lineconstraints,sinceinboth
examplesthedeclarationoftheconstraintisincludedwiththecolumndefinition,
beforethecommathatendsthiscolumnspecificationandenablesasubsequent
columnspecification.
Here’sanotherexampleofin-lineconstraints.Thisexamplecreatesatable
withaNOTNULLconstraint:
TheresultofthisconstraintistoensurethatavalueforSTATUSmustbe
includedwitheachrowenteredintoVENDORS.Thevaluemightbezeroorany
othersingledigit,butitmustbeprovided—itcannotbeleftout.Itcannotbe
unknowntothedatabase.Inotherwords,itcannotbeNULL.
Hereisthesametablewithanameassignedtotheconstraint:
YoumaycombinemultipleconstraintdeclarationsinasingleCREATE
TABLEstatement,likethis:
EachconstraintspecificationistreatedasastandaloneclauseoftheCREATE
statement.Youcanmixandmatchdefinitiontypes;youcanchoosetonamethe
constraintinsomeclausesandnotinothers.Youcanalsomixandmatchin-line
constraintswithout-of-lineconstraints,towhichweturnourattentionnext.
CREATETABLE:Out-of-LineConstraints
Inadditiontoin-lineconstraints,youmayoptionallydefineaconstraintwithina
CREATETABLEstatementafterthecolumnspecifications.Here’sanexample
ofaPRIMARYKEYdefinedwiththeout-of-linesyntax,inthisexampleonline
4:
Afterthefinalcolumnisdefinedforthetable,weincludeacomma,followed
bythereservedwordsPRIMARYKEY.Noticethattheout-of-linesyntax
requiresthatyouindicatewhichcolumn(orcolumns)isaffectedbythe
constraint.Sinceyou’renot“inline”withthecolumnspecification,the
statementcannotknowwhichcolumnyou’reintendingtoconstrain,unlessyou
specificallyindicateitwithintheclause.
Here’sanout-of-lineexamplethatnamestheconstraint:
Thisexamplegivestheconstraintanamethatwe’vechosen.Asisthecase
within-lineconstraints,anyout-of-lineconstraintinwhichyoudonotprovidea
namewillcausetheOracledatabasetoautomaticallygenerateaconstraintname.
AdditionalWaystoCreateConstraints:ALTERTABLE
TheCREATETABLEstatementisn’ttheonlywaytocreateaconstraint.
ConstraintscanalsobecreatedusingtheALTERTABLEstatement.For
example,youcanfirstcreatethePORTStablelikethis:
Afterward,youcanALTERthetabletoaddaconstraintbymodifyingthe
definitionforacolumn.
Intheprecedingcode,we’remodifyingthedeclarationofthecolumnitselfby
addingtheprimarykeyandlettingthesystemassignaname.Thissyntaxresults
inthesameCONSTRAINTasifwehadusedthisinstead:
Inaddition,youcanusethedatadictionarytodeterminewhatthesystem-
generatedconstraintnameis.
Onceyoulearnthesystem-generatednameoftheconstraint(let’ssayit’s
SYS_C0011505),thenyoucanrenameitifyouwish,likethis:
Thosearethein-lineequivalentsofALTERTABLE.Herearetheout-of-line
equivalents.First,hereisanALTERTABLEthataddsaconstraintwithout
specifyinganame:
Alternatively,youcannametheconstraintyourself.
You’llworkwiththedatadictionaryinChapter12.
Warning:NOTNULLIsDifferent
We’reabouttolookatthefivedifferenttypesofconstraints,butbeforewedo,a
wordofwarningaboutthesyntaxvariationsforoneparticularconstraint:the
NOTNULLconstraintisabitdifferentwhenitcomestosyntax.TheNOT
NULLconstraintcannotbecreatedoutofline.Inotherwords,thisisinvalid:
Thisisalsoinvalid:
Eitherofthosewillproduceerrormessagesifyoutrythem.Notableor
constraintwillbecreated.Andyet,thesamesyntaxisperfectlyfineforother
typesofconstraints.Forexample,here’stheUNIQUEconstraint:
Andofcourse,thisisfine,too—thisisthePRIMARYKEYconstraint:
Soremember,NOTNULLcannotbedeclaredwiththe“out-of-line”format.
Theotherscan.Butwait,there’smoreaboutNOTNULL.Thiswon’twork
either:
Andthiswon’teither:
Thosewon’tworkbecausetheyaretheALTERTABLEequivalentsforout-
of-linedeclarations.ButtheALTERTABLEin-lineequivalentsarefine.
Andthisisalsofine:
Thelesson:beware.NOTNULLisabitunusual.It’savalidconstraintand
canbecreatedusingtheotherformsofsyntaxbutnotwiththeout-of-line
format.
Thelistofcertificationobjectivesspecificallystatesthatyouwillbe
testedonthetopicofcreatingconstraintsatthetimeoftablecreation.
Thiscanincludeanyoneofthein-lineorout-of-linecombinationsfor
anyconstraint.Thesyntaxistricky.Studyitthoroughlyandknowitwell.
TheTypesofCONSTRAINTS
Severaltypesofconstraintscanbeappliedtoatable.TheyarePRIMARYKEY,
FOREIGNKEY,NOTNULL,CHECK,andUNIQUE.(NotethattheREFtype
isnotincludedontheexam.)Let’sexploreeachoftheseindetail.
NOTNULL
TheNOTNULLconstraintissimple—whenappliedtoacolumn,itensuresthat
foranyrowthatisaddedtotheTABLE,thecolumnonwhichtheNOTNULL
constraintisappliedwillalwaysbeprovidedwithavalue.Meanwhile,the
column’sdatatypeensuresthatthedataenteredintothecolumnisconsistent
withthedatatype’srules.
Tofullyappreciatethisconstraint,ithelpstounderstandwhattheconceptof
NULLis.Let’stakealook.
TheConceptofNULLThereservedwordNULLis,inmyever-so-humble
opinion,oneofthemostmisunderstoodaspectsoftheSQLdatabase.The
definitionofNULListhe“absenceofinformation.”Sometimesit’s
mischaracterizedas“zero”or“blank,”butthatisincorrect—a“zero,”afterall,is
aknownquantity.Soisablank.ButNULLis“unknown.”It’sthewaythe
databaseacknowledgesthatit,thedatabase,ismerelyanimperfectrepositoryof
informationaboutsomereal-lifebusinesssituationorsomeotherapplicationout
thereintheworld,andthedatabaseisultimatelydependentonthedatait’sbeen
giventomirrorthatreal-lifesituation.Butultimately,it’sreallife—notthe
database—thatisthefinalauthority.So,it’sentirelypossible—andquitelikely
—thatsomeinformationhasn’tbeenprovidedtothedatabase.NULLisa
placeholderforwherethatinformationgoes,specifically,wherethedatabasehas
notbeengivenclearinstructionaboutwhetheravalueexists,andifitdoesexist,
whatvalueitmightbe.
Forexample,considerthefollowingtablecontainingthenamesofcustomers:
Let’saddarowtothistable.
NoticethatnovalueisprovidedhereforMIDDLE_NAME.Whenthis
happens,thenintheabsenceofotherinformation,thedatabasestoresaNULLin
itsplace.
Doesthismeanthatsomewhereoutthereintheworldisareal-lifeperson
namedAngelinaEllisonwhohasnomiddlename?Well,maybeshedoes,and
maybeshedoesn’t.Thepointisthatthedatabasedoesn’tknowwhethershehas
amiddlenameornot—thevalueisunknowntothedatabase.
Thisbecomesimportantwhenitcomestosuchsituationsasmathematical
expressions.Considerthefollowingexpression:
Let’ssaythatthevalueforCRUISE_PRICEis300,andthevaluefor
DISCOUNTisNULLinthedatabase.Inotherwords,maybethere’sadiscount
value,andmaybethereisn’t.Thedatabasedoesn’tknoweitherway—it’s
unknowntothedatabase.ThevalueisNULL.Ifthat’sthecase,thenwhatisthe
answertotheequationgiveninthisexpression?
Theanswerfortheequationof300timesNULLis...what?
Giveup?
TheanswerisNULL.Thereasonissimple:300multipliedby“Idon’tknow”
resultsin...“Idon’tknow.”
Inotherwords,perhapstheequationreallydoeshaveananswer,butifyou
don’tknowwhattheDISCOUNTvalueis,thenyoudon’thaveenough
informationtocalculatetheansweroftheexpression.Therefore,theansweris
unknown—inotherwords,NULL.
We’lladdressNULLsomemoreinthesectiononfunctions.Fornow,allthat
we’reconcernedwithistheNOTNULLconstraint.WhentheNOTNULL
constraintisappliedtoacolumn,you’rerequiringthatanyrowsaddedtothe
tableincludeavalueforthatcolumn.
Let’sgobacktotheearlierexample.IfwehadappliedaNOTNULL
constrainttotheMIDDLE_NAMEcolumnoftheCUSTOMERStable,thenwe
neverwouldhavehadarowlike“AngelinaEllison”withnomiddlename.The
NOTNULLconstraint,ifappliedtotheMIDDLE_NAMEcolumn,wouldhave
requiredamiddlenamevalueforanyrowbeingaddedtothetable.
Intherealworld,therereallyarepeoplewithoutamiddlename.Insuch
situations,anapplicationissmarttousetheNMNorNoMiddleNameentry
toconfirmthatthereisnomiddlenameandthusavoidaNULLvalue,which
isambiguous.
Bydefault,allcolumnsallowNULLvalueswhenfirstcreated.Youmust
choosetoapplyaNOTNULLconstraint—oranequivalent—toacolumnifyou
wanttorequiredataforthatcolumn.
WhenIsay“oranequivalent,”Imeanthattherearealternativewaysto
requiredatainacolumn.OnewayisthePRIMARYKEYconstraint,whichisa
NOTNULLrulecombinedwiththeUNIQUEconstraint,whichyou’llseenext.
IfyouapplyaPRIMARYKEYconstraintonacolumn,youdonotneedto
alsoapplyaNOTNULLconstraint.ButifthereisnoPRIMARYKEY
constraintforagivencolumn,theNOTNULLconstraintwillensurethatany
rowsaddedtothetablewillincludeavalueforthatparticularcolumn.
UNIQUE
TheUNIQUEconstraint,whenappliedtoacolumn,ensuresthatanydataadded
tothecolumninthefuturewillbeuniquewhencomparedtodataalready
existinginthecolumn.Nootherrowwillpossessthesamevalueforthat
particularcolumn.
ThefollowingareafewnotesaboutUNIQUE:
AsingleUNIQUEconstraintcanbeappliedtoonecolumnorto
multiplecolumnsincombination.You’lllearnmoreaboutthisinafew
sentences.
UNIQUE,byitself,allowsNULLvaluestobeaddedtothecolumn.
Itonlyrestrictsdatathat’sprovidedforthecolumntobeingoneofakind
forthecolumn.
NotethatthePRIMARYKEYconstraintrepresentsthecombinationofNOT
NULLandUNIQUE.UsethePRIMARYKEYconstraintinsteadoftheNOT
NULLandUNIQUEconstraintsifyourintentistocreateasingleunique
identifierforeachrowinthetable.
CompositeUNIQUEConstraintYoumaycreateaUNIQUEconstraintthat
appliestomultiplecolumnssimultaneously.Thishastheeffectofrequiringthe
combinationofcolumnstobeunique.Inotherwords,eachindividualcolumn
mayrepeatdata,butcollectivelythecombinationofdatainallofthecolumns
foranygivenrowwillhavetobeunique.
PRIMARYKEY
ThePRIMARYKEYdefinesoneormorecolumnsinatablethatwillformthe
uniqueidentifierforeachrowofdatathatisaddedtothetable.ThePRIMARY
KEYconstraintisacombinationoftheNOTNULLandUNIQUEconstraints.
AtablemayhaveonlyonePRIMARYKEYconstraint.
Asingle-columnPRIMARYKEYisthemostcommonform,anditensures
thatforallrowsofdataaddedtothetableinthefuture,thecolumnuponwhich
thePRIMARYKEYconstrainthasbeenappliedwillalwayscontainavalueand
thatvaluewillalwaysbeuniquewhencomparedtoexistingvaluesthatare
alreadyinthetableforthatparticularcolumn.
HereisanexampleofaCREATETABLEstatementthatcreatesaPRIMARY
KEYconstraint:
Intheprecedingexample,wecreateaPRIMARYKEYconstraintonthe
EMPLOYEE_IDcolumn.Inthisexample,we’vegiventheconstraintanameof
EMPLOYEES_PK.(ThePKsuffixisnotrequired;itispresentedhereasjust
oneofmanygooddesignapproachesthatclarifiestoanyonewhomightreviewa
longlistofdatabaseconstraintslaterthatthisparticularconstraintisaprimary
key.)Nowthatwe’vecreatedthistablewiththePRIMARYKEYconstraint,any
rowthat’saddedtotheEMPLOYEEStableinthefuturewillrequireaunique
valueforeachrowadded.
CompositePrimaryKeysAmulticolumnPRIMARYKEYisbasedontwoor
morecolumnsthatcollectivelyservethesamepurposeasasingle-column
PRIMARYKEY.Inotherwords,thecombinationofcolumnvalueswill
collectivelyhavetobeunique,andallcolumns—individuallyandcollectively—
willhavetocontainvalues.
SeeFigure2-4forsomesampledata.NoticethethreecolumnsCATEGORY,
YEAR,andTICKET.Individuallyeachshowsdatathatrepeatsthroughoutthe
datalisting.Buttogethereachrowofcombinedcolumnsrepresentsuniquedata.
FIGURE2-4 SampledatafromHELP_DESKtable
Here’sanexampleofaCREATETABLEstatementthatwouldcreatea
compositePRIMARYKEYconstrainttosupportFigure2-4:
TheprecedingcodehastheeffectofcreatingNOTNULLandUNIQUE
constraintsacrossallthreecolumns.ForeachrowenteredinthetableHelpDesk,
avaluewillberequiredineachofthethreecolumnsHD_Category,HD_Year,
andHD_Ticket_No,andthecombinationofthosethreevalueswillneedtobe
uniqueforeveryrow.Asyousawintheearliersampledata,it’spossibleto
repeatvaluesintheindividualcolumns,butthecombinationmustalwaysbe
uniqueineachrow.
FOREIGNKEY
AFOREIGNKEYconstraintappliestooneormorecolumnsinaparticular
tableandworksinconjunctionwithareferredtable’sPRIMARYKEY
constraint.AFOREIGNKEYisthefeaturethathelpsensurethattwotablescan
“relate”toeachother,andinmanywaysthisreallyrepresentsthe“heartand
soul”ofwhatarelationaldatabaseisallabout.
TheFOREIGNKEYconstraintdoesthefollowing:
Itidentifiesoneormorecolumnsinthecurrenttable;let’scallthis
thechildtable.
Foreachofthosecolumns,italsoidentifiesoneormore
correspondingcolumnsinasecondtable;let’scallthistheparenttable.
ItensuresthattheparenttablealreadyhaseitheraPRIMARYKEY
orUNIQUEconstraintonthosecorrespondingcolumns.
ItthenensuresthatanyfuturevaluesaddedtotheFOREIGNKEY–
constrainedcolumnsofthechildtablearealreadystoredinthe
correspondingcolumnsoftheparenttable.
Inotherwords,aFOREIGNKEYconstraintonthechildtable,alongwiththe
PRIMARYKEYconstraintontheparenttable,enforcesreferentialintegrity
betweenthetwotables.Thismeansthattheconstraintsworktoensurethatany
futuredatathatisaddedtooneorbothofthetablescontinuestosupportthe
abilitytorelatedatafromonetabletoanother.
Note:ThereferencedtableisnotactuallyrequiredtohaveaPRIMARYKEY
constraintonthereferencedcolumnsbutonlyaUNIQUEconstraintonthe
referencedcolumns.Butyou’llrecallthataPRIMARYKEYconstraintisa
combinationoftheUNIQUEandNOTNULLconstraints,sothePRIMARY
KEYsatisfiestherequirementforaUNIQUEconstraint.
Let’slookatasamplescenario.First,hereisalistingofdatainthePORTS
table:
Next,hereisalistingofinformationintheSHIPStable:
Asyoumighthavealreadysurmised,thevalueforeachship’s
HOME_PORT_IDvalueshouldcorrespondtoaPORT_IDvalueinthePORTS
table.
Toensurethatthetwotablesonlyacceptdatathatsupportsthisbusinessrule
(inotherwords,thatrequiresallHOME_PORT_IDvaluestobevalidPORT_ID
values),youcancreateaPRIMARYKEYconstraintonthePORTStable(ora
UNIQUEconstraint)andthenaFOREIGNKEYconstraintontheSHIPStable
thatcorrelatesbacktothePRIMARYKEYconstraintonthePORTStable.
First,hereisthePORTStable(linenumbersadded):
Next,hereistheSHIPStable:
NotethattheforeignkeyconstraintclauseintheCREATETABLESHIPS
statementstartsonline11andcontinuesthroughline12.Itreferencesthe
PORTStableandthePORTStable’sPORT_IDcolumn,whichalreadyhasa
PRIMARYKEYconstraintappliedtoit.Ifitdidnotalreadyhaveeithera
PRIMARYKEYconstraintoraUNIQUEconstraintonit,thentheCREATE
TABLESHIPSstatementwouldresultinanerrorandletyouknowthatthe
PORTStablealreadymustexistandmusthaveaPRIMARYKEYorUNIQUE
constraintonthePORT_IDcolumn.
TheFOREIGNKEYonSHIPSmakessurethatanyrowaddedtotheSHIPS
tablewillacceptvaluesforHOME_PORT_IDonlyifthatvaluealreadyexistsin
thePORTStable.NotethattheHOME_PORT_IDvalueisnotrequired;ifyour
goalistoensurethattheHOME_PORT_IDvalueisalwaysprovided,you’ll
havetoalsoaddaNOTNULLconstraintonHOME_PORT_IDaswellas
FOREIGNKEY.Thisisonewaytodothat:
Intheprecedingexample,wecreatetwoseparateconstraints.Ofthosetwo
constraints,oneofthemisonlines11through12andexiststomakethe
HOME_PORT_IDcolumnaforeignkey,andtheotherconstraintisattheendof
line10andensuresthatthereisalwaysavalueenteredforHOME_PORT_ID.
Inmyprofessionalexperience,Ifinditmucheasiertocreateforeignkeys
withALTERTABLEstatementsinsteadofwithCREATETABLEstatements.
Onereasonisthattheresultingcodeismoremodular.Notethatyoucannot
createaforeignkeyconstraintthatreferstoatablethatdoesn’texist.Ifyour
goalistobuildascriptforcreating(andifnecessary,re-creating)yourentire
databaseandthenyoutrytocreateallofyourtableswithforeignkeyconstraints
withinyourCREATETABLEstatementssothattheyoccurafterthecreationof
theirrespectiveprimarykeytables,allofthatcansuddenlyturnintoquitea
puzzleoftryingtoensureyourCREATETABLEstatementsallruninthe
correctordersothatyourreferredtablesalreadyexistbeforeyoucreatetables
withforeignkeyconstraintsagainstthoseprimarykeytables.Notonlyissuch
aneffortdifficult,itmayprovetobeimpossibleinadatamodelwherethe
relationshipsruninbothdirections.Allofthisisanunnecessaryeffortwhenyou
caneasilybuildallofyourtablesfirst,thenalterthosetablestoaddforeignkeys
withinaseriesofALTERTABLEstatements,andfinallyplacethemallafter
yourCREATETABLEstatementsasawhole.Thecomplexityjustdescribed
(the“puzzle”)iseliminatedwiththatapproach.Othercomplicationsyoumight
avoidincludethefactthatyoucannotcreateaforeignkeywithinaCREATE
TABLEstatementthatusesthe“asquery”approach—thatapproachcreatesthe
tableandpopulatesitwithdatausingasubquery’sSELECTstatementallat
once.Inthatsortofsituation,theclausetocreateaforeignkeyconstraintisn’t
allowed.Givensuchrestrictions,Ifindnobenefittostrugglingtobuildaforeign
keyfromwithintheCREATETABLEstatementandwouldjustassoonusean
ALTERTABLEstatementwheretheserestrictionsdon’texist.But...that’sjust
me.
ONDELETETheforeignkey/primarykeyrelationshipbetweentwotables
introducessomeadditionalconsiderations.Whatshouldbethebehaviorwhen
rowsareremovedfromtheparenttableforwhichthereisachildrowinachild
table?
UsingtheearlierexampleofPORTSandSHIPS,aSHIPSrowmightcontain
avalueforPORT_ID,indicatingthehomeportoftheship.Butwhatifthecruise
lineceasestodobusinesswithaparticularportandremovestherowforthat
portfromthePORTStable?
OneoptionwouldbetomanuallyeditSHIPSandremoveallreferencestothe
PORT_IDvaluefortheformerport.
Analternative(andbetter)approachistoincludetheONDELETEclause
whenwecreatetheFOREIGNKEYconstraint.
ThepreviousCREATETABLESHIPSstatementdiffersfromtheearlier
exampleintwoways:
TheHOME_PORT_IDcolumn(line10)isnolongerconstrainedby
NOTNULLsincewewanttheoptionofhavingaNULLvaluefor
HOME_PORT_ID.
TheforeignkeyconstraintnowincludesthekeywordsON
DELETESETNULLaspartoftheconstraintdefinition(notethatitis
withintheclosingparenthesesthatendthecreationoftheconstraintbut
notthecreationofthetableSHIPS).
Now,whenaparentrowinPORTSisremoved,theOracledatabasewill
automaticallyinspectrowsinSHIPStoseewhetheranySHIPspecifiedthe
removedport’sprimarykeyintheHOME_PORT_IDcolumn.Ifyes,anyrows
withavaluefortheremovedPORT_IDwillhavethosePORT_IDvalues
replacedwithaNULLvalue.
Asanalternative,youcouldchoosetoremovetheSHIPSrowcompletelyifa
correspondingparentrowinPORTSisremoved.
ThisoptionwillremoveallSHIProwsassignedtoaPORTrowuponthe
PORTrow’sremovalfromthePORTStable.
CHECK
ACHECKconstraintattachesanexpressiontoaconstraint.Inotherwords,it
appliesasmallbitofcodetodefineaparticularbusinessruleonincomingrows
ofdata.ACHECKconstraintmay,forexample,restrictincomingdatasothatall
incomingvaluesarerequiredtobegreaterthansomeminimumvalueorfall
withinasetofpredeterminedoptionalvalues.ACHECKconstraintcanensure
thatatwo-charactercolumnacceptsonlyvalidabbreviationsforAmerican
states,forexample,orthatthedateenteredinonecolumninatableisalways
greaterthanthedateenteredinanothercolumninthesametable.
Here’sanexampleofaCHECKconstraintthatonlyallowsrowsinthe
VENDORStablewithaSTATUSvalueofeither4or5:
WhilerowsmaybeaddedtoVENDORSwithnoSTATUSvalue,theycanbe
givenaSTATUSvalueonlyifitiseither4or5.
AnyvalidSQLexpressionmaybeusedinaCHECKconstraint,withsome
limitations.TheCHECKconditioncannotincludereferencestothefollowing:
Columnsinothertables(notethatothercolumnsinthesametable
areaccepted)
PseudocolumnsCURRVAL,NEXTVAL,LEVEL,orROWNUM
Subqueriesandscalarsubqueryexpressions
User-definedfunctions
Certainfunctionswhosevaluecannotbeknownatthetimeofthe
call:SYSDATE,SYSTIMESTAMP,CURRENT_DATE,
CURRENT_TIMESTAMP,DBTIMEZONE,LOCALTIMESTAMP,
SESSIONTIMEZONE,UID,USER,andUSERENV
Forarowofdatatobeacceptedasinputtogiventable,anyCHECK
constraintpresentmustevaluatetoeitherTRUEorunknown,duetoaNULL.
MultipleConstraints
Atablemaybedeclaredwithmultipleconstraints.Here’sanexample:
Intheprecedingexample,wehaveasingleCREATETABLEstatementthat
createsatablealongwithfiveconstraints:
Auser-namedPRIMARYKEYonVENDOR_ID
Asystem-namedNOTNULLconstraintonVENDOR_NAME
Auser-namedNOTNULLconstraintonSTATUS
TwoCHECKconstraints:oneonSTATUSandanotheron
CATEGORY
AnysingletablemayhaveonlyonePRIMARYKEYconstraint.Itcanhave
anyothercombinationofanyotherconstraints.
FullList
Table2-4displaysafullsetofsampleconstraintstatementsforeachconstraint
typeandeachstatementformat.
TABLE2-4 ConstraintReview
TakeparticularnoteoftheFOREIGNKEYexamples.Themostimportant
portionoftheFOREIGNKEYconstraintdeclarationisnotthekeywords
FOREIGNKEYbutthekeywordREFERENCES.ThetermFOREIGNKEY
appearsinonlyoneformatofthestatement.
DataTypeRestrictions
Therearesomerestrictionsonsomeconstraints.SeeTable2-5forasummaryof
datatyperestrictionsonconstraints.Theserestrictionsmeanthatthedatatypes
identifiedcannotandwillnotreceiveaconstraintappliedagainstthemifthey
areofthetypesindicatedinthetablewitha“No”intheappropriatefield.For
example,thePRIMARYKEYconstraintcannotincludeanycolumnswithadata
typeofBLOB,CLOB,orTIMESTAMPWITHTIMEZONE.(Note,however,
thatconstraintsmaybeappliedtocolumnsthathavethedatatypeof
TIMESTAMPWITHLOCALTIMEZONE.)
TABLE2-5 DataTypesandConstraintRestrictions(No=NotAllowed)
CERTIFICATIONOBJECTIVE2.06
DropColumnsandSetColumnUNUSED
Thissectionlooksatwhatyoucandowhenyoufindthatyouhavecolumnsyou
nolongerwishtokeepinatable.Youcanchoosetodropacolumnfromthe
databaseor,asanalternative,youmayrenderitUNUSED.Thissectionlooksat
howtodoeitherofthesetasks.
DroppingColumns
Alteringatabletodropcolumnsthatarenolongerusedcanfreeupstorage
spaceandpotentiallyimproveperformanceofthetable.Anyconstraintsor
indicesonthecolumnwillalsobedroppedwhenyoudropthecolumn.
Herearetwoexamples—onewithoutparenthesesaroundthecolumnname,
anotherwiththeparentheses:
BotharevalidstatementstodroptheCRUISE_ORDER_DATEcolumn.Note
thatthekeywordCOLUMNisrequiredinthefirstvariation,wheretheDROP
clausesyntaxomitstheparentheses.Inthesecondvariation,thekeyword
COLUMNisomitted,andtheparenthesesareused.
ThefirstvariationislimitedtodroppingonecolumnperSQLstatement.
Usingthesecondvariation,however,youcandropmultiplecolumnsbyusing
thekeywordDROPonetime,omittingthekeywordCOLUMN,andthen
includingapairofparentheses,followedbyalistofthecolumnnamesyouwish
todrop.Forexample:
Youcannotdropallofthecolumnsinatable.Atablemusthaveatleastone
column.
Restrictions
Ifacolumnisreferencedbyaforeignkeyconstraintinanothertable,thenthe
precedingsyntaxwilltriggerawarningmessageandtheattempttodropthe
columnwillfail.
Forexample,considerthiscode:
TheseSQLstatementswillcreatetwotables.CRUISE_ORDERSiscreated
withaPRIMARYKEYconstraint,andORDER_RETURNSiscreatedwitha
FOREIGNKEYconstraintthatreferencestheprimarykeyof
CRUISE_ORDERS.SeeFigure2-5forthedatamodelrepresentingthe
relationship.
FIGURE2-5 DatamodelforCRUISE_ORDERSandORDER_RETURNS
IntheCRUISE_ORDERStable,wecannotsimplydropthe
CRUISE_ORDER_IDcolumn:
InordertodroptheCRUISE_ORDER_IDcolumnweneedtoaddressthe
constraint:
ThestatementaboveworkswhenweaddthekeywordsCASCADE
CONSTRAINTStoindicatethatwewanttodropthecolumnandalsodropany
constraintsonthecolumnaswell.
WhenusingthekeywordsCASCADECONSTRAINTS,notethatboth
CONSTRAINTandCONSTRAINTSworkequallywell.Oracledocumentation
specifiesthekeywordsintheplural:CASCADECONSTRAINTS.
Thereasonwe’repreventedfromdroppingtheCRUISE_ORDER_IDcolumn
inCRUISE_ORDERSisbecauseaforeignkeyconstraintelsewhereis
dependingonthatcolumn.However,thereferencingtabledoesnothavethe
samerestrictionondroppingcolumns.Forexample,thiswillworksuccessfully
withouttheCASCADECONTRAINTSkeywords:
TheDROPCOLUMNstatementexecutessuccessfully,andwhileitmaynot
bereadilyapparent,theFOREIGNKEYconstraintisalsodroppedasaresultof
thisstatement.Afterall,theFOREIGNKEYconstraintisbasedonthe
CRUISE_ORDER_IDcolumninORDER_RETURNS,andsincewe’vedropped
thecolumninthattable,theconstraintisdroppedaswell,withoutthe
CASCADECONSTRAINTSkeywords.However,youmayincludethemifyou
wish:
Thiswouldhaveworkedaswelleitherway.
However,let’slookaslightvariationofthisscenarioandseewhathappens
whenwehaveacompoundkeyasourforeignkey.Seebelow:
Thecodeabovecreatesthesametwotableswesawearlier—
CRUISE_ORDERSandORDER_RETURNS.Asbefore,thereisaforeignkey
constraintonORDER_RETURNSreferencingCRUISE_ORDERS,butnowitis
acompoundforeignkey—thisforeignkeyisbasedontwocolumns:
CRUISE_ORDER_IDandORDER_DATE.
IfwetrytodroponeofthosecolumnsinORDER_RETURNS,wegetan
errormessage:
Onealternativehereistodropbothforeignkeycolumnsatonce:
Alternatively,wecouldstaywithouroriginalplantodroponecolumn,but
accountfortheconstraintintheprocess.IfweaddthekeywordsCASCADE
CONSTRAINTStotheDROPCOLUMNstatementonjustthe
CRUISE_ORDER_IDcolumn,thestatementwillexecutesuccessfully.See
Figure2-6foranexampleoftheresults.
FIGURE2-6 DROPCOLUMNwithCASCADECONSTRAINTS
TheeffectoftheCASCADECONSTRAINTkeywordswiththeDROP
COLUMNclauseistodropthecolumnandalsotodroptheassociated
constraintsonthecolumnthatmightpreventthecolumnfrombeingdropped.
Anadditionalnoteaboutdroppingcolumns:youcannotdropthefinal
remainingcolumninatable;atleastonecolumnmustremaininthetable.
UNUSED
Insteadofdroppingatablecolumnyouarenolongerusing,youmayelectto
declareitunusedandleaveitinplace.OnceyousetacolumnasUNUSED,itis
neveragainavailable;itisasthoughithasbeendropped.Aswithdropped
columns,anyconstraintsorindicesonthecolumnwillalsobedropped.Youwill
neverbeabletorecoveracolumnthatissettoUNUSED.AROLLBACK
statementwillhavenoeffect—anUNUSEDcolumnwillnotberecoveredina
ROLLBACKoperation.OnceacolumnissettoUNUSED,youcanaddnew
columnsthathavethesamenameasanyunusedcolumnsforthetable.
OnceacolumnhasbeensettoUNUSED,itcanneverberecovered.
Sowhywouldn’tyoujustDROPacolumninsteadofsettingittoUNUSED?
OnereasonistheperformancefortheDROPstatementversustheSET
UNUSEDapproach.Ifyou’reworkingwithaverylargetableorsetoftables,
andyouneedtodropsomecolumns,youmayfindthatthesystemperformance
forexecutingtheDROPistemporarilyunacceptable,particularlyforasystem
thatisinheavyproduction.Ifthisisanissue,andyouneedtoachievethelook-
and-feelofadroppedcolumnimmediately,thentheSETUNUSEDalternativeis
awelcomeoption.Theperformanceisspeedy,theresultsare—forallpractical
purposes—thesame,andyoucanalwaysscheduleatimelatertocomebackand
dropthecolumnduringaperiodoflowactivityinthedatabase.
Onethingtokeepinmind:there’salimittothetotalnumberofcolumnsany
giventablecanhave.Thatlimitis1,000—youcannotcreatemorethana
thousandcolumnsinanyonetable.IfyousetacolumntobeUNUSED,that
columnwillstillcountaspartofthethousandcolumnstowardyourlimit,until
youeventuallyDROPthecolumn—whichyoucando;we’lldiscusshowtodrop
anunusedcolumninabit.
ThesyntaxforSETUNUSEDisvirtuallyidenticaltotheALTERTABLE...
DROPsyntax.SimplyreplaceDROPwiththekeywordsSETUNUSED,andthe
restisthesame.Forexample:
Note:intheaboveexample,ifthecolumnCRUISE_ORDER_DATEdoesn’t
alreadyexist,thestatementwillfailwithanerrormessage.Andaswesawwith
DROP,ifthecolumndoesexistandalsoisconstrained—ifthereisanexisting
constraintonthecolumn—theALTERwillfail.
AswithDROP,thesyntaxforchangingmultiplecolumnstotheUNUSED
staterequiresparenthesesandeliminatestheCOLUMNreservedword,likeso:
YoucansetasmanycolumnstoUNUSEDasyouwish.Theonlyrequirement
isthatyoumust,asyoumightguess,satisfyallconstraintsandother
requirementsofatableanditsstructure—forexample,thetablestillmusthave
atleastonevalidcolumnatanytime—soyoucannotsetallofitscolumnsto
UNUSED.
TablesthathaveanycolumnsthataresettoUNUSEDcanbefoundinthe
datadictionaryviewUSER_UNUSED_COL_TABS.However,thisviewdoesn’t
revealanycolumnnamesthatareunused;itsimplygivesyouthenamesofany
andalltablesthatcontainunusedcolumns,andanumericcountofhowmany
unusedcolumnseachonecontains.Youcannotrecovertheunusedcolumns,nor
canyouevenidentifythem.Butyoucandropthem.Todropthoseunused
columns,usethisstatement:
Forexample:
Thisstatementwilldropallunusedcolumnsthatareassociatedwiththetable
ORDER_RETURNS.
CERTIFICATIONOBJECTIVE2.07
CreateandUseExternalTables
Anexternaltableisaread-onlytablethatisdefinedwithinthedatabasebut
existsoutsideofthedatabase.Inmoretechnicalterms,theexternaltable’s
metadataisstoredinsidethedatabase,andthedataitcontainsisoutsideofthe
database.
Externaltableshaveanumberofrestrictionsonthem.Youcanquerythem
withtheSELECTstatement,butyoucannotuseanyotherDMLstatementson
them.Youcan’tcreateanINDEXonthem,andtheywon’tacceptconstraints.
Benefits
Sowhywouldyoucreateanexternaltable?Theirprimarybenefitistocreatean
easy-to-usebridgebetweenSQLtablesandnon-databasedatasources.Ifyou’ve
everusedtheOracletoolSQL*Loader,orDataPump,thenyou’llbepleasedto
discoverthattheexternaltablefeaturewasdesignedtoincorporatethe
functionalityfoundinthosetoolsintoaSQLcontext.
Agreatexampleiswhenyouhavesomenon-SQLdatasourcethatregularly
producesinformationneededinthedatabase,suchasaflatfiletransfer,aweb
sitereference,aspreadsheetapplication,alegacy3GLapplication,orsomething
comparable.Ifthatdatasourceiscapableofprovidingsomesortofformatted
flatfile,thenitcanbestructuredinsuchawaythatitcanbecopieddirectlyinto
afilethattheSQLexternaltablewillinstantlyrecognizeandbeabletoquery.In
otherwords,itwillcreateasortofone-waydatatransferintotheSQLdatabase,
butusingSQLSELECTstatementsinsteadofutilitiessuchasSQL*Loader.
CreatingExternalTables
Tocreateanexternaltable,youcandeclareitscolumnsandtheirdatatypes.You
canalsopopulatetheexternaltablewithasubqueryatthetimeyoucreateit.
Butthat’saboutallyoucandowithexternaltables.Theyarerestrictedina
numberofways:
YoucannotcreateacolumnwithaLOBdatatype—noCLOB,
BLOB,NCLOB,etc.
Youcannotaddaconstrainttoanexternaltable.
YoucannotchangethecolumnofanexternaltabletoUNUSED.If
youtry,SQLwillprocessthestatementbutwillactuallydropthe
column.
Essentially,allyoudowithanexternaltableisdeclareitsstructureanddefine
theparametersbywhichtheSQLdatabasecommunicateswiththeexternaltable.
Inordertoestablishthatcommunication,youmustfirstunderstandtwosubjects:
DIRECTORYobjects
TheOracleutilitiesSQL*LoaderandOracleDataPump
We’lllookatthosenext,andthenwe’llcreateanexternaltable.
DIRECTORYObjects
Tocreateanexternaltable,we’llneedtoidentifythelocationintheoperating
systemwheretheexternalfilecontainingthetablewillreside.Forthis,weneed
tolookattheCREATEDIRECTORYstatement.
TheCREATEDIRECTORYstatementcreatesanobjectinthedatabasethat
representsthenameofadirectoryontheserver’sfilesystem.Here’san
example:
CREATEORREPLACEDIRECTORYdirectory_nameAS
directory_reference;
wheredirectory_nameisanameyouspecify,justasyouwouldanyother
databaseobject,anddirectory_referenceisastringliteral,surroundedbysingle
quotationmarks,thatidentifiesalocationwithinyourOracleserver’sfile
system,intowhichyouwishforexternaltablestobestored.Forexample:
CREATEORREPLACEDIRECTORYBANK_FILESAS
'F:\bnk_files\trnsfr';
Theresultofthisstatementisthatwe’vejustcreatedanobjectinthedatabase
namedBANK_FILESthatlookstotheoperatingsystemwheretheOracleserver
residesandassumesthatthedirectoryreferenceinthestringliteralisconsistent
withthesyntaxrequiredforthatparticularoperatingsystem.Inthiscase,we’re
pointingtoaWindowsdrive“F:”anditsrootleveldirectory“bnk_files”,within
whichisthesubdirectory“trnsfr”—thatsubdirectoryisourtarget.
TheDIRECTORYobjectwillnotparsethisreferencebutinsteadwilljust
storeitasis.Ifit’sincorrect,youwon’tfindoutuntillaterwhenyoutrytouse
theDIRECTORYobject.
Also,theDIRECTORYobjectwillnotcreatethesubdirectory;the
assumptionhereisthatthesubdirectoryalreadyexists.Ifitdoesnot,youwon’t
getanerrormessageuntilyouusetheDIRECTORYobjectlater.
ThekeywordsORREPLACEareoptional.
Inourexample,thenameBANK_FILESisanamewespecify.Thisnameis
thenameassignedtotheobject,andthisnameishowwewillreferencethe
DIRECTORYobjectinthefuture.
Onceadirectoryhasbeencreated,theownermustgrantREADand/or
WRITEaccesstoanyuserwhomayuseit:
GRANTREADONDIRECTORYdirectory_nameTOusername;
Thatincludesuserswhomaywishtouseexternaltablesthatarebuiltwiththe
directoryobjects.
OracleUtilities
TheOracledatabaseprovidesanumberofutilitiesthataccompanytheir
databaseproduct.Thoseutilitiesthatareimportanttoexternaltablesinclude
SQL*Loader
OracleDataPumpExport
OracleDataPumpImport
EachisdocumentedintheOracleCorporationreferencemanualtitled
“OracleUtilities”.Together,theutilitiesprovidecapabilitiesthatallowexternal
datasourcestocommunicatewithSQLobjectswithinthedatabase.
Acompletereviewoftheircapabilitiesisbeyondthescopeoftheexamand
thereforethisbook.Butit’simportantfortheexamtorecognizethatalarge
componentofthedefinitionsassociatedwiththedeclarationofanexternaltable
comesfromtheseutilities.
CreatinganExternalTable
Let’swalkthroughanexample.Let’ssaywehaveanexternaltextfilecontaining
thefollowingdataaboutinvoices:
Wewanttocreateanexternaltableforthisdata;let’scallit
INVOICE_DATA.TXT.
First,wegotothefilesystemonwhichtheOracledatabaseresides,locatethe
samedrive,andwecreateasubdirectoryoffoftherootlevel.We’llcallit
“LOAD_INVOICES”.ThenwecreatetheassociatedDIRECTORYobject:
Bythispoint,wewon’tnecessarilyhavetohavecreatedthe
LOAD_INVOICESdirectory,nortohaveputtheINVOICE_DATA.TXTfilein
thatdirectory.Butforthesakeofourexample,nowwedoso,beforecontinuing.
Next,weexecuteaCREATETABLEstatementthatreferencesthedirectory,
alongwiththenecessaryclausestotellOracleSQLtoloadtheexternalfile,and
howtoloadit:
Oncethisstatementexecutes,weendupwithanexternaltableinthedatabase
calledINVOICES_EXTERNAL.
Notelines2through4wherewedeclaredourtableusingthedata
typesCHAR.You’llrecallthesearefixed-lengthdatatypes.Wedidthis
toaccommodatethetransferofrowsinfromthetextfileinlines12
through14.Eachcolumn’sdatatypeissettoCHAR,thefixed-length
alphanumericdatatype,andthecountsforeachdatatypecorrespondto
thecountsofthecolumnsinthetextfile‘INVOICE_DATA.TXT’,which
isidentifiedinline16andisinthedirectorystoredinthedirectoryobject
INVOICE_FILES,namedinline8.
Lines1through5formacompleteCREATETABLEstatementby
themselves,withouttheexternaltableclause.Butstartingonline6are
thekeywordsandclausesusedtodeclaretheexternaltable,andtogether,
lines1through17formthecompleteCREATETABLEstatementforour
example.
Line6includesthekeywordsORGANIZATIONEXTERNAL,
whicharerequired.
Line7iswherewespecifythatweareusingORACLE_LOADER,
akatheSQL*Loaderfeatures.AnalternativeTYPEvalueherewouldbe
ORACLE_DATAPUMP.
Line9beginsthesetofvaluesforACCESSPARAMETERS,which
areenclosedwithintheparenthesesthatopenonline10andcloseonline
15.
ThreeACCESSPARAMETERSareusedhere:RECORDS,SKIP,
andFIELDS.
Line10—RECORDSDELIMITEDBYNEWLINE—meansthat
eachnewlinestartsanewrowofdatafortheINVOICES_EXTERNAL
table.
Line11—SKIP2—tellsORACLE_LOADERthatthefirsttwolines
oftheINVOICE_DATA.TXTfilearetobeskipped—theyjustcontain
headerinformation.
Line12—FIELDS—startsthespecificationsforeachcolumn,where
eachcolumn’slengthiscarefullyspecifiedtomatchthelengthinthe
INVOICES_DATA.TXTfile.
ManymoreACCESSPARAMETERSexistthatarenotinvokedhere.Icould
probablywriteaseparatebookjustonalltheoptionsandfeaturesthatexistwith
thevariousclausesforthetypesORACLE_LOADERand
ORACLE_DATAPUMP.ButIwon’t,andyoushouldn’tneedthatfortheexam.
UsinganExternalTable
Oncewe’vecreatedanexternaltable,wecanSELECTfromitjustlikeanyother
table.Forexample:
Manyexternaltableswillstartwithsourcedatathatisroughandunformatted.
However,thefirststepisjusttogetitinthedatabase.Oncethatisaccomplished,
youcanusethevariousconversionfunctionsandotherfeaturesofSQLtoclean
upandreformatthedata:
Notetheoutput—thenumbersarereformatted,thedatevaluesareconverted,
theaccountnumbershavebeentrimmedup,andeverythinglooksterrific.
Externaltablescanbequeriedlikeanytableorviewinthedatabase.
RememberthatyoucannotuseINSERT,UPDATE,orDELETEstatements
onexternaltables.
CERTIFICATIONSUMMARY
Themaindatabaseobjectsthataresubjectsoftheexamincludetables,views,
sequences,synonyms,indexes,constraints,users,androles.Atablestoresdata.
Aviewissomethingthatlooksandactslikeatablebutservesasafilterontoone
ormoretables.Asequenceisacounter,andit’softenusedtogenerateunique
numbersforstoringidentifierswithnewrowsthatareaddedtoatable.A
synonymisanaliasforanotherobject.Anindexisanobjectthatprovides
lookupsupporttoatableinordertospeedupqueriesonthetable.Aconstraintis
aruleonatablethatcontrolswhatsortofdatacanbestoredinthetable.Auser
isanobjectthatdefinesauseraccount.Arolerepresentsasetofoneormore
privilegesthataregrantedtoauserinorderforthatusertohaveaccessrightsto
otherobjects.
Alldatabaseobjectsareeitherschemaornonschemaobjects.Schemaobjects
areownedbyauserandexistwithinauseraccount.Nonschemaobjectsexistto
supportthedatabaseatlarge.Ofthemaindatabaseobjectsyouarelookingatfor
theexam,theschemaobjectsaretable,view,sequence,privatesynonym,index,
andconstraint.Thenonschemaobjectsareuser,role,andpublicsynonym.
YouusetheCREATETABLEstatementtocreateatable,namethetable,
namethecolumns,assigndatatypestothecolumns,andoptionallycreate
variousconstraintstothetableaswell.
Objectsthatexistinthesamenamespacemusthaveuniquenames.Objects
thatexistindifferentnamespacesmayhaveduplicatenames.Indexeshavetheir
ownnamespacewithinaschema;sodoconstraints.Beyondthat,aschemahas
onenamespaceforthecollectivesetoftables,views,sequences,andprivate
synonyms.Outsideoftheschema,userandroleobjectsshareanamespacefor
theentiredatabase;publicsynonymshavetheirownnamespacefortheentire
database.
Columnsmustbeassignedadatatypewhentheyarecreated.Datatypes
includecharacter,numeric,anddatedatatypes.Characterdatatypesinclude
CHARandVARCHAR2;numericdatatypesincludeNUMBERandFLOAT;
datedatatypesincludeDATE,TIMESTAMP,TIMESTAMPWITHTIME
ZONE,TIMESTAMPWITHLOCALTIMEZONE,INTERVALYEARTO
MONTH,andINTERVALDAYTOSECOND.Therearealsolargeobject,or
LOB,datatypes,suchasBLOB,CLOB,andNCLOB.
ConstraintscanbecreatedwithintheCREATETABLEstatementor
afterward,intheALTERTABLEstatement.Theycanbecreatedinline,
meaningaspartofacolumn’sdefinition,oroutofline,meaningafterwardasa
separatelineitemwithintheCREATETABLEorALTERTABLEstatement.An
exceptionisNOTNULL,whichcannotbecreatedoutofline.
ThefivetypesofconstraintsareNOTNULL,UNIQUE,PRIMARYKEY,
FOREIGNKEY,andCHECK.
TheTRUNCATETABLEstatementcanbeusedinthosespecial
circumstanceswhereyoumaywanttoremovealltherowsofatablewithout
firinganyDMLtriggersorselectivelyremovingdatafromanyassociated
INDEXobjects,aswouldhappenwithaDELETEstatement.Insuchasituation,
TRUNCATETABLEisanattractivealternativetodroppingandre-creatinga
table,sinceitleavesthetableanditsrelationshipsintact,aswellasanyindexes.
TRUNCATETABLEcanremovestatementsmuchmorequicklywithless
processingoverheadthanDELETE.BecauseTRUNCATETABLEfiresno
DMLtriggers,itisconsideredDDL;therefore,itsuseperformsanimplicit
commit,anditsresultscannotberolledback.
IfTRUNCATETABLEisappliedtoatableforwhichthereexistsan
associatedparenttablewithaforeignkeyusingtheONCASCADEDELETE
clause,thenTRUNCATETABLEwillresultinanerrorifthereareinfactrows
inthatparenttablethatmightotherwisebedeletedasaresultofONCASCADE
DELETE.Forthissituation,youmustuseTRUNCATETABLE...CASCADE
toconfirmyourdesiretoremovethoseparentrows.
Columnscanbedroppedfromatable.Columnsthatarepartofareferential
constraint—inotherwords,aFOREIGNKEYconstraint—maybedroppedwith
theCASCADECONSTRAINTSkeywordsaddedsothattheconstraintsare
droppedaswell.ColumnsmaybesettoUNUSEDtorenderthemvirtually
dropped,butsparingtheprocessingoverheadrequiredtodropacolumn.An
UNUSEDcolumnisnolongeravailable,anditcanbedroppedlater.
Externaltablesexistoutsideofthedatabase.Theprocessthatsupportsthem
hasitsrootsandsyntaxintheSQL*Loadertool,buttheresultisatablethatcan
bedescribedjustlikeanyother.Externaltablescanbequeriedbutcannot
receivedatainputviaINSERT,UPDATE,orDELETE.
✓TWO-MINUTEDRILL
CategorizetheMainDatabaseObjects
Tablesstoredata.
Viewsserveasasortof“window”ontotables.
Indexesprovidelookupsupporttospeedqueriesonatable,likean
indextoabook.
Sequencesaresimplecounterobjects.
Synonymsarealternativenamesforexistingobjects.
Constraintsarerulesontables.
Usersareobjectsthatownothernonuserobjects.
Rolesaresetsofrights,orprivileges,thatcanbegrantedtoauser
togivethatuseraccesstootherobjects.
Objectsareeitherschemaornonschemaobjects.
Tables,views,indexes,sequences,andprivatesynonymsare
schemaobjects.
Usersandroles,alongwithpublicsynonyms,arenonschema
objects.
CreateaSimpleTable
TheCREATETABLEstatementisusedtocreateatable.
Youassignanametoatablebyusingtherulesofnamingdatabase
objects.
Youalsoassignnamestothetable’scolumnsusingthesamerules.
Alltableshaveatleastonecolumn.
ReviewtheTableStructure
TheDESCcommandcanbeusedtodisplayatable’sstructure.
Thestructureincludesthetablename,tablecolumns,datatypes,
andoptionalconstraints.
ListtheDataTypesThatAreAvailableforColumns
Eachcolumnmustbeassignedadatatype.
Datatypesincludenumeric,character,anddatetypes,suchas
VARCHAR2,NUMBER,andDATE.
Datatypesalsoincludelargeobjecttypes,includingBLOB.
ExplainHowConstraintsAreCreatedattheTimeofTable
Creation
ThetypesofconstraintsareNOTNULL,UNIQUE,PRIMARY
KEY,FOREIGNKEY,andCHECK.
AcolumnwithaNOTNULLconstraintmustbeassignedavalue
foreachrowthatisaddedtothetable.
AUNIQUEconstraintrequiresthatifdataisaddedtoacolumnfor
agivenrow,thatdatamustbeuniqueforanyexistingvaluealreadyin
thecolumn.
APRIMARYKEYconstraintisthecombinationofNOTNULL
andUNIQUE.
APRIMARYKEYmaybeassignedtoonecolumnora
combinationofcolumns.
APRIMARYKEYassignedtoacombinationofcolumnsiscalleda
compositekey.
AsingletablemayhaveonlyonePRIMARYKEY.
AFOREIGNKEYcorrelatesoneormorecolumnsinonetablewith
asetofsimilarcolumnsinasecondtable.
AFOREIGNKEYrequiresthatthesecondtablealreadyhaveeither
aPRIMARYKEYconstraintoraUNIQUEconstraintassignedtothe
correlatedcolumnsbeforetheFOREIGNKEYcanbecreated.
Oncecreated,theFOREIGNKEYensuresthatanyvaluesaddedto
thetablewillmatchexistingvaluesinthePRIMARYKEYcolumnsof
thesecondtable.
ConstraintscanbecreatedwiththeCREATETABLEstatementor
withintheALTERTABLEstatement.
Constraintscanbedefinedaspartofthecolumndefinitions(inline)
orafter(outofline).OneexceptionisthatNOTNULLcanbecreatedin
linebutnotoutofline.
DropColumnsandSetColumnUNUSED
TheDROPclauseofALTERTABLEcanbeusedtoremovea
columnfromatable.
OnceacolumnisremovedwithDROP,thedataislost.
Droppingacolumncanconsumesignificantprocessingtimeifthe
tableinvolvedcontainsalotofdata.
Ifyoudropacolumnwithaconstraint,theconstraintisalso
dropped.Thesameistrueforanyindexobjectsonthecolumn;theyare
alsodropped.
SETUNUSEDrendersacolumnpermanentlyunavailable;itcannot
berecovered.
TheSETUNUSEDclausecanbenefitalargetableinheavy
productionthatcannotaffordtheoverheadprocessingpowerofaDROP.
AfteratablehascolumnsthataresettoUNUSED,theycanbe
droppedwiththeDROPUNUSEDCOLUMNSclauseofALTER
TABLE.
CreateandUseExternalTables
Anexternaltableisaread-onlytablewithinthedatabasethatstores
dataoutsideofthedatabase.
Thecommunicationbetweentheexternaltable’sdatastoragefile
anddatabaseobjectsisbasedonthelogicofSQL*LoaderorOracle
DataPump.
YouuseadatabaseobjectknownastheDIRECTORYobjectaspart
ofthedefinitionofanexternaltable.
SELFTEST
Thefollowingquestionswillhelpyoumeasureyourunderstandingofthe
materialpresentedinthischapter.Chooseonecorrectanswerforeachquestion
unlessotherwisedirected.
CategorizetheMainDatabaseObjects
1.Atableiswhichofthefollowing?(Chooseallthatapply.)
A.Aschemaobject
B.Anonschemaobject
C.Arole
D.Alloftheabove
2.Whichofthefollowingareschemaobjects?(Chooseallthatapply.)
A.SEQUENCE
B.PASSWORD
C.INDEX
D.ROLE
3.ACONSTRAINTisassignedtowhichofthefollowing?(Chooseall
thatapply.)
A.TABLE
B.SYNONYM
C.SEQUENCE
D.INDEX
CreateaSimpleTable
4.WhichofthefollowingarevalidCREATETABLEstatements?
(Choosethree.)
5.Whichofthefollowingoptionscanbeusedwiththereservedword
CREATEtoformthebeginningofacompleteSQLstatement?(Choose
three.)
A.TABLE
B.VIEW
C.CONSTRAINT
D.SEQUENCE
6.YouareloggedintouserFINANCE.Itiscurrentlytheonlyschemain
theentiredatabase.Thefollowingexistinthedatabase:
–AVIEWnamedVENDORS
–ACONSTRAINTnamedVENDORS
–AnINDEXnamedCUSTOMER#ADDRESS
YouattempttoexecutethefollowingSQLstatement:
Whichoneofthefollowingistrue?
A.ThequestionisflawedbecauseyoucannothaveanINDEX
namedCUSTOMER#ADDRESS.
B.ThequestionisflawedbecauseyoucannothaveaVIEWanda
CONSTRAINTwithidenticalnamesinthesameschema.
C.TheSQLstatementwillfailtoexecuteandresultinanerror
messagebecauseyoucannotcreateaTABLEnamewiththe#
character.
D.TheSQLstatementwillfailtoexecuteandresultinanerror
messagebecauseyoucannotcreateaTABLEthathasthesamename
asanINDEXinthesameschema.
E.TheSQLstatementwillexecute,andtheTABLEwillbecreated.
7.Youhaveasingledatabase,withonlyoneschema.Thefollowingfour
objectsexistinthedatabase:
–ATABLEnamedPRODUCT_CATALOG
–ATABLEnamedADS
–AUSERnamedPRODUCT_CATALOG
–AVIEWnamedCONFERENCE_SCHEDULE
Howmanyofthefourobjectsareownedbytheschema?
A.0
B.2
C.3
D.4
8.WhichofthefollowingistrueaboutROLES?
A.Rolesareschemaobjectsbutonlywhencreatedfromwithina
useraccount.
B.RolesareinthesamenamespaceasCONSTRAINTS.
C.RolesareinthesamenamespaceasTABLES.
D.RolesareinthesamenamespaceasUSERS.
ReviewtheTableStructure
9.TheDESCcommandcanbeusedtodowhichofthefollowing?
A.Showatable’scolumnsandthedatatypesofthosecolumns
B.Showabriefparagraphdescribingwhatthetabledoes
C.Showatable’snameandwhocreatedit
D.Showthedatathatiscontainedwithinatable
ListtheDataTypesThatAreAvailableforColumns
10.YouattempttoexecutethefollowingSQLstatement:
Whichoneofthefollowingistrue?
A.Theexecutionfailsbecausethereisnoprecisionindicatedfor
NUMBER.
B.Theexecutionfailsbecausethereisnoprecisionindicatedfor
VARCHAR2.
C.Theexecutionfailsbecausethereisnoprecisionindicatedfor
CHAR.
D.Theexecutionsucceeds,andthetableiscreated.
11.ThefollowingSQLstatementscreateatablewithacolumnnamedA,then
addarowtothattable,thenquerythetable:
WhatisthedisplayedoutputoftheSELECTstatement?
A.3.1415
B.3.142
C.3.141
D.Noneoftheabove
ExplainHowConstraintsAreCreatedattheTimeofTable
Creation
12.WhichofthefollowingSQLstatementscreatesatablethatwillreject
attemptstoINSERTarowwithNULLvaluesenteredintothe
POSITION_IDcolumn?
13.ReviewthefollowingSQLstatement:
AssumethereisnotablealreadycalledSHIPPING_ORDERinthe
database.Whatwillbetheresultofanattempttoexecutethepreceding
SQLstatement?
A.ThestatementwillfailbecausethedatatypeforORDER_YEAR
isaCHAR,andCHARdatatypesaren’tallowedinaPRIMARY
KEYconstraint.
B.Thestatementwillfailbecausethereisnoprecisionforthe
ORDER_IDcolumn’sdatatype.
C.Thetablewillbecreated,buttheprimarykeyconstraintwillnot
becreatedbecausethenamedoesnotincludethe_PKsuffix.
D.Thestatementwillsucceed:thetablewillbecreated,andthe
primarykeywillalsobecreated.
14.ReviewthefollowingSQLstatement:
AssumethereisnotablealreadycalledPERSONNELinthedatabase.
WhatwillbetheresultofanattempttoexecutetheprecedingSQL
statement?
A.Thestatementwillfailbecauseyoucannotcreatetwoprimary
keyconstraintsonthetable.
B.Thestatementwillsuccessfullycreatethetableandthefirst
primarykeybutnotthesecond.
C.Thestatementwillsuccessfullycreateasingletableandone
compositeprimarykeyconsistingoftwocolumns.
D.Thestatementwillsuccessfullycreatethetableandtwoprimary
keys.
DropColumnsandSetColumnUNUSED
15.ThedifferencebetweendroppingacolumnfromatablewithDROPand
settingacolumntobeUNUSEDis:
A.AnUNUSEDcolumncanberecovered.
B.TheUNUSEDcolumnanditsdataareretainedwithinthetable’s
storageallocationandcountsagainstthetotallimitonthenumberof
columnsthetableisallowedtohave.
C.AcolumnthatisdroppedwithDROPnolongerappearswithin
thetable’sdescriptionasshownwiththeDESCorDESCRIBE
statement,whereasacolumnthatissettoUNUSEDstillappearsin
thetable’sstructureasshownintheoutputoftheDESCstatement.
D.Nothing.
CreateandUseExternalTables
16.ThepurposeoftheCREATEDIRECTORYstatementistocreateanamed
objectinthedatabase:
A.Thatlistsnamesofuseraccountsthathaveexternalprivileges
B.Thatcontainslookupreferencematerialforqueries
C.ThatidentifiestherootdirectoryoftheOracleserverinstallation
D.Thatpointstoadirectoryyouchoosesomewherewithinthe
Oracleserver’sfilesystem
SELFTESTANSWERS
CategorizetheMainDatabaseObjects
1. A.Alldatabaseobjectsareeitherschemaornonschemaobjects,and
atablefallsunderthecategoryofschemaobjects.
B,C,andDareincorrect.Atable,whichisaschemaobject,isnota
nonschemaobject.Aroleisanotherformofadatabaseobject.
2. AandC.Asequenceandanindexarebothschemaobjects.They
areownedbyaschema.
BandDareincorrect.Apasswordisnotadatabaseobject.Arole
consistsofoneormoreprivilegesthatareassignedtoauserobject,and
boththeuserandroleobjectsarenonschemaobjects.
3. A.ACONSTRAINTisarulethatrestrictswhatsortofdatacanbe
addedtoatable.
B,C,andDareincorrect.Youcannotattachaconstrainttoasynonym,
sequence,orindexobject.
CreateaSimpleTable
4. B,C,andD.Underscoresareacceptablecharactersinanydatabase
tablename,providedthatifthenameisunquoted,thefirstcharacterisnot
anunderscore.Quotationmarks,whenused,enableanycharactertobe
usedthroughoutthename,includingspaces,althoughallfuturereferences
tothenamewillbecase-sensitiveandrequirequotationmarks.Finally,
mixedcasecanbeusedtocreateatablename,althoughwhencreated,the
tablenamewillbestoredandtreatedasuppercase.Allfuturereferencesto
thenamemaycontinuetobeinanycase—SQLwillperformthenecessary
caseconversionsothatthetablename,onapracticallevel,istreatedas
thoughitiscase-insensitive,eventhoughitisstoredinternallyin
uppercase.
Aisincorrect.Thefirstcharacterinanydatabaseobjectnamemustbea
letter;itcannotbeanumberorspecialcharacter—unlessthenameis
enclosedindoublequotationmarks,whichthiswasnot.
5. A,B,andD.ThedatabaseobjectsTABLE,VIEW,andSEQUENCE
caneachbecreateddirectlywithaCREATEreservedwordtoforma
completeSQLstatement.
Cisincorrect.ACONSTRAINTisnotcreateddirectlywithaCREATE
CONSTRAINTstatementbutinsteadiscreatedindirectlyasaclausewithin
theCREATETABLEandALTERTABLEstatements.
6. E.Thetablenamemayincludethe#character,aslongasitisn’tthe
firstcharacterinthetablename.AnINDEXinthesameschemaisallowed
tohavethesamenamesincetheINDEXisinsideitsownnamespacewithin
theschema,separatefromthenamespaceinwhichtheTABLEwillbe
created.
A,B,C,andDareincorrect.Thequestionisnotflawed—thehashmark
(#)isanacceptablecharacterinanobjectnameanywherefromthesecond
characterpositionthroughtothefinalcharacter.Youareallowedtohavea
VIEWandaCONSTRAINTwiththesamenameswithinasingleschema,
sinceCONSTRAINTSarecontainedwithintheirownuniquenamespacein
eachschema.
7. C.TABLEandVIEWobjectsareschemaobjects,andsinceyou
haveonlyoneschemainthedatabase,thenbothhavetobeownedbythe
onlyschemainthedatabase.ButtheUSERobjectisanonschemaobject.
Infact,inthisone-schemadatabaseitmustbetheschemathatownsthe
otherobjects.Itdoesnotownitself,anditexistsatthedatabaselevel.
A,B,andDareincorrect.
8. D.BothROLESandUSERSexistatthedatabaselevelandshare
thesamenamespace.
A,B,andCareincorrect.Itdoesn’tmatterthataROLEiscreatedfrom
withinaschema’suseraccount;it’sstillanonschemaobjectandexistsin
thesamenamespace.CONSTRAINTSandTABLESbothhavetheir
namespaceswithinaschema.
ReviewtheTableStructure
9. A.DESC,orDESCRIBE,presentsadisplayshowingatable’s
columnsandthedatatypesofthosecolumns.
B,C,andDareincorrect.TheDESCcommanddoesn’tshowdata,the
creator,oratextdescriptionofthetable.Allofthatinformationisavailable
throughothermeans,nottheDESCcommand.
ListtheDataTypesThatAreAvailableforColumns
10. B.TheVARCHAR2datatyperequiresprecision,forexample,
VARCHAR2(30).
A,C,andDareincorrect.NUMBERandCHARcanbedeclared
withoutprecision.ButVARCHAR2cannot,andthestatementwillfail.
11. B.TheNUMBERdatatypehasaprecisionof5andascaleof3.Thescale
indicatesthatthreedigits—butnomorethanthreedigits—totherightof
thedecimalpointareallowed.Thenumberisroundedoff,and5isalways
roundedup.
A,C,andDareincorrect.
ExplainHowConstraintsAreCreatedattheTimeofTable
Creation
12. B.Theprimarykeyconstraintperformstwomainjobs,oneofwhichisto
ensurethatthecolumnuponwhichitisappliedisneverallowedtobe
NULL.Theotherjobistoensurethatanyvalueaddedtothatcolumnis
UNIQUE.
A,C,andDareincorrect.TheUNIQUEconstraintallowsforNULL
values.ThereisnosuchthingasaREQUIREDconstraint.ThereisaNOT
NULLconstraintthatessentiallydoeswhataREQUIREDconstraintmight
do,ifoneexisted.However,thisparticularformofconstraintcreationis
theout-of-lineform,andtheNOTNULLconstraintcannotbecreatedwith
theout-of-lineform,butratherthein-lineform.
13. D.Thesyntaxofthestatementisfine.Boththetableandtheprimarykey
constraintwillbesuccessfullycreated.
A,B,andCareincorrect.Itisperfectlyacceptabletocreateaprimary
keyinanyform—single-columnorcomposite—withtheCHARdatatype.
NUMBERisalsofinewithorwithoutprecisionand/orscale.The_PK
suffixisnotrequired.
14. A.Thestatementisattemptingtocreatetwodifferentprimarykey
constraints,butatablemayhaveonlyoneprimarykeyandnomore.The
statementwillfailwithanerror.
B,C,andDareincorrect.Thesyntaxisnotattemptingacomposite
primarykeybutrathertwoseparateprimarykeyconstraints,andthatisnot
allowedonanytable.Theentirestatementwillfailtoexecute.
DropColumnsandSetColumnUNUSED
15. B.TheUNUSEDcolumnisstillstoredaspartofthetable.There’sno
storagebenefittothetable—nospaceisreclaimedfromanunusedcolumn.
A,C,andDareincorrect.NeitheracolumndroppedwithDROPnoran
UNUSEDcolumncanbeseeninthetable’sstructure.Acolumn,onceset
toUNUSED,canneverberecovered.Itcanonlybedropped.
CreateandUseExternalTables
16. D.CREATEDIRECTORYletsyoucreateadatabaseobjectnamefora
directoryyouchoose.Lateryoucanusethisobjectforcreatinganexternal
table.
A,B,andCareincorrect.
3
ManipulatingData
T
CERTIFICATIONOBJECTIVES
3.01TruncateData
3.02InsertRowsintoaTable
3.03UpdateRowsinaTable
3.04DeleteRowsfromaTable
3.05ControlTransactions
✓Two-MinuteDrill
Q&ASelfTest
hischapterbeginstolookatthatpartofSQLknownasDataManipulation
Language.We’llgetsomeperspectivebylookingatDMLwithregardto
whereitfitsintothelargercontextofSQL.ThenwewillreviewDML
statementsandtheirusageandlookatsomespecificexamples.Finallywewill
reviewsomesupplementalstatementsthatareusedtocontroltransactions
involvingDML.
CERTIFICATIONOBJECTIVE3.01
TruncateData
Wheneveryouneedtoremoveoneormorerowsfromatable,theDELETE
statementisgenerallyyourbestchoice.ButDELETEmaynotbeidealinevery
situation.Forexample,let’ssayyourtableisequippedwithindexestospeedup
queryingandwithtriggerstoprotectthelogicalintegrityofthedatainthetable.
Thoseareallgreat,butifyourtablealsohasalargenumberofrows,thena
DELETEstatementwillinvokerow-by-rowprocessingtoreviewandupdate
eachindexandfireallappropriatetriggersonarow-by-rowbasis.There’s
nothingwrongwiththat,ofcourse;infact,there’severythingrightaboutit.This
iswhatawell-designeddatabasemightneedtodo,dependingonitsbusiness
requirements.
Butincertaincircumstanceswhereyousimplyneedtoremovealltherowsin
thetable,youmayfindthatallthatdetailedprocessingisunnecessary;youmay
needtosimplyremoveallthedata.
Insuchascenario,youmightconsiderthealternativeofdroppingthetable
andthencreatingitagain.Butifyouweretodothat,youwouldalsohavetore-
createthetable’sindexesandtriggers.Dependingonthetable’srelationships
withothertables,theextraworkrequiredtomakethisallhappencouldgrow
considerably.Inanattempttobypasssomeunnecessaryprocessingofa
DELETEstatement,youmaybeinadvertentlycreatingagoodamountof
otherwiseunnecessaryrework.YetDELETEmayitselfimpactperformanceona
largeproductiondatabase.
ThisiswhereTRUNCATETABLEcomesin.TheTRUNCATETABLE
statementdoesthefollowing:
Removesalltherowsinagiventable
Removesalldataintheassociatedindexes
FiresnoDMLtriggers
ByavoidingtheDROPTABLE/CREATETABLEapproach,using
TRUNCATETABLEdoesthefollowing:
Leavesthetableandindexstructuresintact
Leavesalldependenciesintact,suchaschildtables
TRUNCATETABLEalsodoesthefollowing:
DoesnotleverageundospacelikeDELETEwoulddo
Performsanimplicitcommit
RequirestheDROP_ANY_TABLEprivilege
Becauseoftheimplicitcommit,TRUNCATETABLEdoesthefollowing:
Cannotberolledback
DoesnotworkwithFLASHBACK_TABLE
IsconsideredDDL
ThefactthatTRUNCATETABLEisDDLiswhyTRUNCATETABLEdoes
notfireanyDMLtriggers,suchasONDELETE,whichdoesnotapply.
Thefollowingisanexampleofastatementthatwilltruncatethetable
VENDORS:
TRUNCATETABLEVENDORS;
ThekeywordTRUNCATEisfollowedbytherequiredwordTABLE,which
isfollowedbythenameofthetable.(ThekeywordTABLEisrequiredbecause
TRUNCATEmayalsobeusedtotruncateanotherobjectcalledacluster,which
isnotontheexam,sowearenotconcernedwithithere.)Thisuseof
TRUNCATETABLEwillremoveallthedatafromtheVENDORStable,aswell
asitsindexes,andautomaticallyperformacommittothedatabase.
RecursivelyTruncateChildTables
Oracle12coffersanewfeaturewithTRUNCATETABLE,whichistheoption
toCASCADEthetruncationtodependentchildtables:
TRUNCATETABLEVENDORSCASCADE;
Earlierwelookedathowaparent-childtablerelationshipcanbecreatedwith
aforeignkeyconstraintwiththeoptionalONDELETEclause,whichtellsthe
databasetotakespecificactiononachildrowifandwhenaparentrowis
removedfromtheparenttable.TheTRUNCATETABLEstatement’s
CASCADEworksinasimilarfashion.
Let’slookagainattheexamplesofthetablesPORTSandSHIPS(line
numbersadded):
NotetheONDELETECASCADEclauseonline14.Nowlet’saddsome
rowstothetables:
We’vecreatedaportforAtlanta,andashipisassignedtoAtlanta.Butthen
someonepointsoutthatAtlanta,whileabeautifulcity,islandlocked,soperhaps
weshouldn’ttrytoportashipthere.ClearlyweneedtoremoveAtlantafromthe
PORTStable.
WecouldsimplyDELETEtherow:
TheDELETEstatementwillremovealltherowsfromPORTS,andrightnow
thereisonlytheonerowfortheAtlantaport.Butitwillalsodeleteanyrowsfor
SHIPSassignedtotheAtlantaport,thankstotheONDELETECASCADE
clauseincludedintheSHIPStable’sforeignkeyconstraint(seeline14inthe
previouscodelisting).
If,insteadofDELETE,weneededtouseTRUNCATETABLE,thenwe
couldtrythis,butitwillnotwork:
WiththepresenceofrowsinPORTSforwhichtherearecorresponding
HOME_PORT_IDvaluesinSHIPSandbecauseSHIPSistiedtoPORTSusing
theONDELETECASCADEclause,thissetsuptheriskthatrowsinSHIPS
mayhaveHOME_PORT_IDvaluesthatcorrespondtoPORT_IDvaluesin
PORTS.Sureenough,thatisthecase,resultinginanerrormessagewiththe
TRUNCATETABLEstatementearlier:“ORA-02266:unique/primarykeysin
tablereferencedbyenabledbyforeignkeys.”TogetTRUNCATETABLEto
workinthissituation,weneedtheCASCADEclause.
WithoutCASCADE,theOracledatabasewillnotcompletetheTRUNCATE
TABLEstatementsincetheONDELETECASCADEclauseinaseparatechild
tablewouldforcetheremovalofadditionalrowsofdatabeyondjustthePORTS
table.TheTRUNCATETABLE…CASCADEstatementtellstheOracle
databasetogoforwardwiththeremovalofthoserowsanyway.
NotethatweneedednosuchadditionalclausewithDELETE.Thepresence
ofONDELETECASCADEcausestheadditionalrowstoberemovedwithno
additionaldirectionfromtheDELETEstatement.ButtheTRUNCATETABLE
statementhastheadditionalrequirementfortheCASCADEclause.Remember,
thisappliesonlytosituationswhereachildtablehastheONDELETE
CASCADEclauseofitsforeignkeyconstraint.
CERTIFICATIONOBJECTIVE3.02
InsertRowsintoaTable
TheSQLstatementtoaddrowstoatableistheINSERTstatement.TheINSERT
statementisoftenusedtoaddonerowatatime,butitmayalsobeusedtoadd
multiplerowsatoncebydrawingthemfromelsewhereinthedatabaseand
addingthemtothetargettablewithasingleINSERTstatement.INSERTmay
useexpressions.
INSERTaddsrowsintoaTABLEobject,butitcanalsobeusedwithcertain
VIEWobjects.However,aswewillsee,aVIEWissimplyafilterontooneor
moretables,soultimatelyINSERTisstilladdingrowstoaTABLEobject.
DefaultColumnList
Let’slookatanexampleofanINSERTstatement.We’regoingtoaddarowto
theCRUISEStable.First,let’sdescribetheCRUISEStablesothatwecansee
thecolumns—we’llneedtoknowthenamesanddatatypesofthosecolumnsso
wecanbuildourINSERTstatement(seeFigure3-1).
FIGURE3-1 TheCRUISEStable
Here’sanexampleofanINSERTstatementyoucoulduseonthistable:
AswithallSQLstatements,theINSERTstatementcanbeonasingleline,or
itcanspanmultiplelines.Thechoicedoesn’tmatterasfarasthesyntaxis
concerned.Weseparatedthecomponentsoftheprecedingstatementinorderto
discussitmoreeasilyhere.Thestatementendswithasemicolon,whichisatthe
endofline8inthisexample.
Let’sanalyzethesyntaxforthisexampleofINSERT:
Line1ThereservedwordsINSERTandINTO,followedbythe
nameofthetargettable.
Lines2–4Withinasetofparentheses,alistofthetable’scolumns.
Theorderinwhichthecolumnsappeardoesnotneedtomatchtheorder
ofcolumnsinthetable’sstructureasshownwiththeDESCcommand,
nordoesitneedtoincludeallofthetable’scolumns,aslongaswe
provideforallofthe“required”columns,inotherwords,thosewitha
NOTNULLconstraintorsomethingcomparable(likethePRIMARY
KEYconstraint).AnycolumnassignedaDEFAULTvaluemaybe
omittedfromthelist.
Line5ThereservedwordVALUES.
Lines6–8Withinasetofparentheses,aseriesofvaluestobeadded
toeachofthecolumnslistedin—inthisexample—lines2–4.Valuesare
expressionsandarelistedinthesameorderaseachvalue’s
correspondingcolumnlistedinlines2–4.Forexample,thefirstvaluein
thisliststartinginline6willbecomethevalueforthefirstcolumn
identifiedonline2,thesecondvalueidentifiedinline6willbeinserted
intothesecondcolumnidentifiedinline2,andsoon.
WhentheINSERTstatementissubmittedforexecution,thefollowingsteps
willbeperformedbeforethestatementreturnsanyresults:
Theexistenceandvalidityofthetableinline1willbeconfirmed.
Theexistenceandvalidityofthecolumnsinlines2–4willbe
confirmed.
Theexpressionsinlines6–8willbeevaluated.
Thedatatypesoftheexpressionsinlines6–8willbecompared
againstthedatatypesoftheassociatedcolumnsandevaluatedfor
compatibility.
Thevaluesoftheexpressionsinlines6–8willbeappliedtoany
constraintsthatmightexistinthetable.
Ifallrequirementsfordataentrytothetargettablearesatisfiedandthe
INSERTstatementisdeterminedtobevalid,thenthestatementwillexecute,and
therowwillbeinsertedintothetable.
Inthisexample,thecolumnsinourINSERTstatement(lines2–4)justhappen
toaccountforeverycolumninthetable’sstructureandinthedefaultsequence
ofhowthecolumnsexistinthetable.Inthissortofsituation,thecolumnlistat
thebeginningoftheINSERTstatementisnotrequired.Inotherwords,wecould
haveomittedlines2–4inthisexample;thefollowingisavalidalternativetothe
INSERTstatementwejustreviewed:
Thisexamplewillproducethesameresultbecausetheexpressionsinlines3–5
justhappentocoincidewiththecolumnsintheCRUISEStable,innumber
(thereareseven)andindatatype,aswesawinFigure3-1.Ifthatstructure
changes,thepreviousstatementmayfail.Forexample,ifanewcolumnisadded
toCRUISES,thepreviousexamplewillfail,whereasthepriorINSERT
example,whichnamesthecolumns,willprobablycontinuefunctioning
correctly.
INSERTmustrespectalldatatypesandconstraintsinthetargettable.
Nowyouaresurelyaskingyourselfwhichisbetter:(a)toidentifythecolumn
listbynameor(b)toleveragethedefaultapproach,omitthelistofcolumn
names,andsimplyaccountforallofthetable’scolumnsinthelistofvalues?
Theanswerisitdependsonthesituation.Generallyspeaking,inmyexperience,
Itendtoprefer(a),whereyounameeachcolumnspecificallyanddonotdepend
onthedefault.Thereareseveraladvantagestothisapproach.Oneadvantageis
thatyourcodewillstillexecutesuccessfullyifthecolumnsinthetableare
alteredinvariousways.Forexample,ifthetableisdroppedandre-createdin
somesortoffutureupgradeormaintenanceeffort,thecolumnscouldendupina
differentsequence.ThatcouldtriggerasyntaxerrorwiththedefaultINSERT,or
itcouldresultinsomethingworse—nosyntaxerrorbutacolumnmismatch
wherethedatatypeshappentolineupandtheINSERTstatementworks
technicallybutisn’tputtingthedatainthecolumnsyouoriginallyintended.
Forexample,considertheSQLstatementsshowninFigure3-2.Herewesee
atablecalledTEST_SCORESandanINSERTstatementforit.Notethatthe
INSERTusesthedefaultcolumnlist.There’snothingwrongwiththat—
technically.ButnowlookatFigure3-3.NoticethattheTEST_SCORES
columnsareinadifferentorder.YetthesameINSERTstatement—withthe
syntaxthatdoesnotlistcolumnsbyname—successfullyexecutes.Why?
BecauseSQLseesonlythedatatypesofthelistofvaluesintheINSERT
statement.Inthiscase,bothvaluesbeinginsertedarenumericliterals,and
numericliteralsareacceptableineithercolumn.So,whatistheintenthere?Is
100thevalueforTEST_SCORE_IDand85thevalueforSCORE?Orisitthe
otherwayaround?Thepointisthatyoucannottellinthisparticularvariationof
INSERTstatementsyntax.
FIGURE3-2 TheTEST_SCOREStable
FIGURE3-3 TheTEST_SCOREStablewithadifferentstructure
Byalwaysenumeratingthelistofcolumnnames,youcanavoidany
confusion.However,thereisoneissuetokeepinmind:if,inthefuture,thetable
intowhichyouareinsertingvaluesmightbemodifiedinsuchawaythatnew
columnsareaddedtoit,thenyou’llneedtoremembertorevisitINSERT
statementslikethisandeditthemifnecessary.Ifyou’veenumeratedacolumn
listandthetableislateralteredwiththeadditionofnewcolumns,thenbarring
anyconstraintsortriggersthatmightyieldastatementfailure,youroldINSERT
statementwillcontinuetofunctionnormally;itjustwon’tprovidedataforthe
newcolumns.ButinalltheprofessionalsituationsI’veencountered,thesetypes
ofissuesarelessproblematicandeasiertoresolvethantheproblemofdatatype
matchesthatsupportillogicaldataentry.Themoralofthisstoryistoalways
identifyyourcolumnnamesinanINSERTstatement.
EnumeratedColumnList
TheINSERTsyntaxwejustreviewedassignsvaluestoeachcolumninthetable.
Itaccomplishesthatfeatbyfirstlisting,byname,eachcolumninthetable,in
thesameorderinwhichthecolumnsappearinthetable’sstructure.Butyouare
notrequiredtolistthecolumnsinorder.Here’sanexample:
ThisisavalidINSERTstatement.Noticehowinlines2–3welistthe
columnsinadifferentorderfromthetablestructure.Thisisfine,solongasthe
listofexpressionstobeinserted(lines5and6)areintheorderasthecolumnlist
(lines2and3).Theyareasfollows:
AnotherchangewiththisINSERTisthatwedonotincludeeverycolumnin
thetable.WeignorethecolumnCRUISE_TYPE_ID.That’salsofine,provided
thatallrequiredcolumnsareincluded,andinthisparticulartable,wehavetwo
columnsthatareNOTNULL,andbothareincludedinthisINSERTstatement.
DataTypeConversion
WhentheINSERTstatementisevaluatedforsyntacticalcorrectness,thedata
typesofthevalueslistedinlines5–6willbecomparedtothedatatypesofthe
columnsidentifiedinlines2–3.Thedatatypesmustbecompatible.Butthe
operativewordhereiscompatible,notidentical.Forexample,thisworks:
Noticethatthe101valueinline2isinquotationmarks,identifyingthatvalue
asastringliteral.Normallythatisusedfortext,andinthiscasethetextisbeing
assignedtothecolumnCAPTAIN_ID,whichacceptsonlynumericdata.Butno
errormessageoccurs,andthisINSERTwillexecute.ThereasonisthatOracle
SQLissmartenoughtofigureoutthatthetextcontainedwithintheliteralvalue
inthissituationhappenstobenumericdata.Thestatementwillexecute
successfully.
ThisfeatureisknowninOracledocumentationasimplicitdatatype
conversion.OracleCorporationformallyadvisesthatsoftwaredevelopersavoid
dependingonimplicitdatatypeconversioninapplicationdevelopmentandrely
insteadonexplicitdatatypeconversion,whichwe’lllookatwhenwereview
SQLconversionfunctionssuchasTO_CHAR,TO_NUMBER,andTO_DATE.
Theruleofthumbisthatwhereveritmakessense,OracleSQLwillperform
animplicitdatatypeconversionifatallpossible.Naturallyitcannotconvert
somethinglike'Montana'toaDATEdatatype.Butifyoutrytoenteranumeric
valuesuchas2011intoadatatypesuchasVARCHAR2,animplicitdatatype
conversionwillconvertthevalueof2011to'2011'andthestatementwill
succeed.
INSERTandConstraints
Ifwehappentoincludedatathatviolatesaconstraint,thenwemightgetarun-
timeerror.Thispointbearsreiteration:violationofaconstraintisarun-time
error.Inotherwords,itisnotasyntaxerror.Forexample,let’ssaywewritethe
followingCREATETABLEstatement:
ThistableincludesaCHECKconstraintthatwilllimitanyvaluesfor
CRUISE_NAMEtooneofthelistedstrings:'Hawaii','Bahamas','Bermuda',
'Mexico',or'DayatSea'.Anythingelsewillberejected.
Next,let’speekaheadalittlebitandcreateaSEQUENCEobject,likethis:
ThisCREATESEQUENCEstatementcreatesanobjectthatwilldispense
individualvalues,andwe’lluseittogenerateprimarykeyvaluesforour
INSERTstatements.
Next,let’susethatSEQUENCEobjectandissuethefollowingINSERT
statement:
ThisINSERTstatementaddsasinglerowtotheCRUISEStable.Thenewrow
consistsoftwovalues.Inline4,thefirstvaluecallsonthenewlycreated
SEQUENCEobjectandasksforthenextavailablevaluefromthesequence,
indicatedwiththereservedwordNEXTVAL.Giventhatoursequenceisnew
anditwascreatedwithallthedefaultsettingsandhasneverbeenusedbefore,
thenthenextavailablevaluewillbe1.Thesecondvaluewe’reinsertinghereis
theliteralstring'Hawaii',whichissyntacticallycorrect,anditalsosatisfiesthe
constraintobjectattachedtotheCRUISE_NAMEcolumnintheCREATE
TABLEstatement.
However,hadweviolatedtheconstraint,theINSERTcouldbesyntactically
correctbutlogicallyincorrect.Forexample,considerthisvariationonthesame
INSERTstatement:
Inthisexample,thestring'HawaiiandBack'violatestheCHECKconstraint
wecreatedfortheCRUISEStable.Therefore,eventhoughthisversionofthe
INSERTstatementissyntacticallycorrect,itwillfailonexecution,andthe
CHECKconstraintwillrejecttheattempttoenterthisrow.SeeFigure3-4fora
sampleoftheexecutionerrorresultingfromthisINSERTstatement—thisis
shownintheSQLDevelopertoolinsteadoftheSQL*Plustoolthatyouhave
seenuptonow.
FIGURE3-4 ExecutionerrorfortheINSERTstatement
NotetheORA-02290run-timeerrormessage.TheCHECKconstraintis
identified,andyoucanseeinthisdisplaythatI’mloggedintothe
EFCODD_TESTuseraccount.
It’salsoworthnotingthateventhoughtheINSERTstatementfailed,the
sequencegeneratoradvancedbyone.Anyattempttoinvokeasequence
generatorwilladvancethenumbercount,eveniftheDMLstatementfails.
We’refarfromdonewithINSERT,andwewilladdressmoreadvanced
conceptsofINSERTinlaterchapters,whenwelookatsubqueries(Chapter9)
andlargedatasets(Chapter13).Butfirst,let’slookattherestofthemajorDML
statements.
CERTIFICATIONOBJECTIVE3.03
UpdateRowsinaTable
TheUPDATEstatementisaDMLstatementthatisusedtomodifyexistingdata
inthedatabase.Itoperatesononetableatatime.Itisusedtomodifyoneor
morecolumnsofdataandcanbeusedtochangealltherowsinatable,oritcan
beusedtochangedatainonlyselectedrows.
AswithINSERT,theUPDATEstatementcanalsoworkwithcertainVIEW
objectstoupdatedataintheTABLEobjectsuponwhichtheVIEWobjectsmay
beconstructed.We’lllookatthismorelaterwhenwediscussVIEWobjects.
Let’slookatanexampleofUPDATE:
ThisUPDATEwilllookforanyandallexistingrowsintheCRUISEStable
wherethevalueforCRUISE_IDequals1.Forallofthoserows,itwillchange
thevalueforCRUISE_NAMEto'Bahamas'andchangethevaluefor
START_DATEto'01-DEC-11'.
WhenwesaytheUPDATEstatementwill“change”thosevalues,this
exampleofUPDATEdoesn’tcareifthecolumninquestionisalreadypopulated
withdataoriftheexistingvalueisNULL.Eitherway,UPDATEwilloverwrite
whateverelsemayormaynotalreadybethereandplacethenewvalueinthe
column.
Let’slookalittlemorecloselyatthesyntaxofthisUPDATEstatement:
Line1ThereservedwordUPDATE,followedbythenameofthe
targettable
Lines2–3OneoccurrenceofthereservedwordSET,followedbya
seriesofoneormoreexpressionsconsistingofthesefourelements:
Acolumnname
Theassignmentoperator(=)
Anexpressionresultinginavaluecompatiblewiththecolumn’s
datatype
Eitheracomma,ifadditionalexpressionsaretofollow,or
nothing,ifthelistofexpressionsiscompleted
Line4AnoptionalWHEREclause,definingaconditiontoidentify
rowsinthetable
Nowthatwe’vereviewedoursampleUPDATEindetail,let’sobserveafew
importantissuesabouttheUPDATEstatementingeneral:
TheseriesofcolumnsenumeratedintheSETclausedoesnotneed
torefertothetable’scolumnsinanyparticularorder.
TheSETclausedoesnotneedtoreferenceallrequiredcolumns—in
otherwords,NOTNULLcolumns.Remember,UPDATEisnotaddinga
rowbutmodifyingexistingdata.AssumingtheUPDATEstatement’s
WHEREclausespecifiesanyexistingoneormorerows,thoserowsare
alreadyinthetable,sopresumablythedatawithineachrowalready
honorsallrequiredconstraints.
AnycolumnnamesomittedfromtheUPDATEstatement’sSET
clausewillnotbechanged.
AnyattemptbyUPDATEtoimplementachangethatviolatesa
constraintwillberejectedandfailtoexecute.
TheWHEREclausespecifiesthecriteriafordeterminingwhich
rowsinthetablewillbeupdated.Ifnosuchrowsarefound,the
UPDATEstatementupdatesnorows,anditexecutessuccessfullywith
noerrors.
TheWHEREclauseisnotrequiredinUPDATE.Ifitisomitted,the
UPDATEwillperformoneachrowinthetable.
Expressions
TheUPDATEstatementcanuseexpressionstoassignvaluestoanygiven
column.Expressionsmayincludeliteralvalues,validcolumnreferences,
mathematicalexpressions,andSQLfunctions.Let’slookatanexampletogetan
ideaofwhatcanbedone.Here’sanUPDATEstatementthatusesexpressionsin
theSETclause:
TheprecedingstatementisanUPDATEstatementintendedtochangedatain
atablecalledCOMPENSATION.TheUPDATEstatementwillchangedatain
anyandallrowswithavalueof83intheEMPLOYEE_NUMBERcolumn.For
anyrowthatmeetsthisWHEREcriteria,UPDATEwillchangethevaluesintwo
columns:
TheSALARYcolumnischangedtoequalitselftimes1.03.Thishas
theeffectofincreasingitsownvalueby3percentor—ifNULL—leaving
thecolumnNULL.
TheLAST_CHANGED_DATEcolumnissettothevalueof
SYSDATE.SYSDATEisabuilt-inSQLfunctionthatcontainsthe
currentdateandtimeaccordingtotheoperatingsystemwhereverthe
databaseisinstalled.
ThisexampledemonstrateshowexpressionscanbeusedwithinanUPDATE
statement.
Constraints
IftheUPDATEstatementviolatesanyconstraintonatable,theentireUPDATE
statementwillberejected,andnoneofthemodificationswillbeacceptedfor
anyoftherows.Inotherwords,iftheUPDATEstatementattemptstochange
anydatainanycolumnofanyrowinatableandanyonechangeresultsinany
oneconstraintviolation,thentheentireUPDATEstatementisrejected.
Forexample,reviewtheseSQLstatements:
Inthiscode,wecreateatablePROJECTSthatincludesacoupleofconstraints;
oneinparticularistheCHECKconstraintthatlimitsanyvalueintheCOST
columntonumbersthatarelessthanamillion.
NowseeFigure3-5.NoticehowweissueanUPDATEstatementinwhichwe
increasethecostofeachprojectby20percent.Thiswillcausetherowidentified
byPROJECT_ID2tobumpupoverourlimitationontheCOSTcolumn.The
resultisthattheentireUPDATEstatementisrejected.
FIGURE3-5 UPDATEstatement:oneconstraintviolationrejectstheentire
statement
However,seeFigure3-6;hereweexecuteaslightvariationofthatUPDATE
statementwhereweavoidtheproblemrow,andtheUPDATEstatement
executes.
FIGURE3-6 UPDATEstatement:noconstraintsviolated
Don’tgettheUPDATEstatement’sSETclausemixedupwiththe
“set”operatorsofUNION,INTERSECT,andMINUS.Thosearetwo
completelyseparateissues,andbothareaddressedontheexam.You’ll
studythesetoperatorsinChapter11.
TheWHEREClause
TheUPDATEstatement’sWHEREclauseisarguablyitsmostpowerfuland
importantfeature.WHEREdetermineswhichoftheexistingrowsinthetable
willbemodifiedbytheSETclause.TheWHEREclauseisnotuniquetothe
UPDATEstatement;WHEREisalsousedforsimilarpurposeswithDELETE
andSELECT.
Awordaboutterminology:UPDATEcanbeusedonlytomodifyexistingrows
inatable.Onapracticallevel,endusersmayspeakintermsof“adding”data
toatablewhenalltheyreallymeanisthattheywanttosetavaluetoacolumn
withinanexistingrow.Inotherwords,whenyouarespeakingwith
nontechnicalusersandtheyspeakof“adding”datatoatable,beawarethat
thisdoesn’tnecessarilymeanyou’llbeusingtheINSERTstatement;an
UPDATEmayactuallybeinorder.Similarly,youcanuseUPDATEto
“remove”valuesfromthetablebysettingagivenrow’scolumntoNULL.But
youaren’tnecessarilyremovingarowinthatsituation—justmodifyingthe
contentsofit.Toremovearow,youneedtheDELETEstatement.Youneedto
keepalerttothedifferencesbetweennaturallanguageandSQLsyntaxonthe
jobandontheexam.
CERTIFICATIONOBJECTIVE3.04
DeleteRowsfromaTable
TheDELETEstatementisusedtoremoverowsfromtablesinthedatabase.
RowsareidentifiedbytheWHEREclauseofDELETE.IfWHEREisomitted,
DELETEremovesalltherowsfromthetable.
WhenDELETEidentifiesarowwiththeWHEREclause,itremovesthe
entirerowfromthetable,notjustanindividualdataelementfromwithinarow.
Ifyourgoalistoremoveasinglevaluefromwithinanexistingrow,while
retainingtherowinthetable,thenyoudon’tuseDELETE;youuseUPDATEto
identifytherowandsetthedesiredvaluetoNULL.
TheDELETEclauseissimple.Here’sanexample:
ThissampledeletesanyandallrowsinthePROJECT_LISTINGtablewherea
columncalledCONSTRUCTION_IDcontainsavalueof12.Allrowsthat
containa12intheCONSTRUCTION_IDcolumnwillbedeletedfromthetable.
Let’slookatthissamplestatement:
Line1TherequiredreservedwordDELETE,followedbythe
optionalreservedwordFROM,followedbytherequirednameofthe
targettable
Line2AnoptionalWHEREclause
AsIjustsaid,thereservedwordFROMisoptional.Inotherwords,this
variationoftheprecedingDELETEstatementisalsovalid:
ThisDELETEstatementperformsthesamefunctionwithoutthereservedword
FROM.
TheWHEREclauseforDELETEperformsthesamefunctionastheWHERE
clauseintheUPDATEstatementandintheSELECTstatement,inthesensethat
itspecifiesrowsforprocessingbytheparticularDMLstatementwithwhichyou
areworking.Butbeware:inthecaseoftheDELETEstatement,ifyouomitthe
WHEREclause,you’lldeleteeveryrowinthetable.
CERTIFICATIONOBJECTIVE3.05
ControlTransactions
Sofarinthischapter,we’velookedattheSQLstatementsINSERT,UPDATE,
andDELETE.Thosethreestatements,alongwithSELECT,formasetofSQL
statementsknownasDataManipulationLanguage.
ThereareothertypesofSQLstatements,butonetypeinparticularisof
specialimportancetoDML.Asmentionedearlierinthechapter,thattypeis
knownasTransactionControlLanguage.Thesestatementsareimportanttoany
SQLsessioninwhichyouuseDMLstatements,asTCLstatementsprovidethe
functionalitytosaveorundothechangesmadewithDMLstatements.
We’lllookatthreeTCLstatementsinthissection:COMMIT,ROLLBACK,
andSAVEPOINT:
COMMITYoucouldthinkofCOMMITasSAVE.Saveschangesto
thedatabasesincethesessionbeganorsincethemostrecentcommit
eventinthesession,whicheverismorerecent.
ROLLBACKUndoeschangestothedatabaseperformedwithina
session,backtothelast“commit”pointinthesession.
SAVEPOINTTemporarilysetsanoptionalpoint,ormarker,ina
sessiontoempowerfuture“commit”or“rollback”actionsbyproviding
oneormoreoptionalpointsatwhichyoumay—ormaynot—undo
changes.
NotethatTCLperformschangesmadetothedatabasewithinasession.A
sessionbeginswhenasingleuserlogsinandcontinuesastheuserengagesina
seriesoftransactions,untiltheuserdisconnectsfromthedatabasebyeither
loggingoutorbreakingtheconnectioninsomefashion.Asingleusermay
engageinmultiplesessions.Buttheoppositeisnottrue:onlyoneusercan
engageinagivensession.ThisdiscussionaboutTCLisconcernedwith
sessions,includingthetransactionsthatoccurwithinagivensessionandhow
theimpactofthosetransactionsmayormaynotimpactothersessions.
Withthatinmind,let’slookateachstatementinmoredetail.
COMMIT
TheSQLstatementCOMMITisusedtosavechangesmadewithinasessionto
anytablesthathavebeenmodifiedbytheDMLstatementsINSERT,UPDATE,
andDELETE.COMMITmakeschangestothedatabasepermanent,andonce
committed,thosechangescannolongerbeundonewithaROLLBACK
statement.Thatisn’ttosaythatthedatacannotbechangedbackwithadditional
DMLstatements;ofcourseitcan.ButbeforeaCOMMITisexecuted,changes
tothedatabasecanbeundonewithaROLLBACKstatement.Afterthe
COMMIT,however,thatoptionnolongerexists.
AseriesofSQLstatementsisconsideredatransactionbySQLandistreated
asoneunit.Thechangesyoumakewithinatransactionarenotmadepermanent
untiltheyarecommitted.Acommiteventcompletesatransaction.
Therearetwokindsofcommitevents:
Anexplicitcommit,whichoccurswhentheCOMMITstatementis
executed
Animplicitcommit,whichoccursautomaticallywhencertain
databaseeventsoccur
Untilacommiteventofeithertypeoccurs,nochangesthatmayhavebeen
performedtotablesinthedatabasearemadepermanent,andallchangeshave
thepotentialforbeingundone.
ExplicitcommitsoccurwhentheCOMMITstatementisexecuted.Implicit
commitsoccurwithouttheCOMMITstatement,butinsteadoccurwhencertain
typesofdatabaseeventsoccur.Let’sdiscussbothsituations.
ExplicitCommit
AnexplicitcommitoccurswhenevertheCOMMITstatementisexecuted.The
syntaxofCOMMITissimple:
COMMIT;
TheCOMMITstatementhasafewparametersthatarenotrequiredforthe
exam.Oneworthnoting,however,istheoptionalkeywordWORK,asfollows:
COMMITWORK;
TheWORKkeywordisincludedforcompliancewithANSIstandardSQL,butit
isnotrequiredinOracleSQL.Itprovidesnoadditionalfunctionalitybeyondthe
COMMITstatementitself.
TounderstandhowCOMMITworks,considerthefollowingseriesofSQL
statements:
Inthisseriesofstatements,thechangetothetablemadebytheINSERT
statementismadepermanentbytheCOMMITstatement.Withoutit,the
INSERTchangescouldbeundonewithaROLLBACKstatement.
ImplicitCommit
Animplicitcommitoccurswhencertaineventstakeplaceinthedatabase.Those
eventsinclude
Immediatelybeforeandimmediatelyafteranattempttoexecuteany
DDLstatement,suchasCREATE,ALTER,DROP,GRANT,or
REVOKE.Note:EveniftheDDLstatementfailswithanexecutionerror
(asopposedtoasyntaxerror),the“before”implicitcommitisexecuted
andtakeseffect.
AnormalexitfrommostofOracle’sutilitiesandtools,suchas
SQL*PlusorSQLDeveloper.(OneexceptionisOracle’sprecompilers,
whichdonotperformanimplicitcommituponexitbutinsteadperforma
rollback.)
Whentheseeventstakeplace,animplicitcommitisautomaticallyexecuted,
meaningthatalluncommittedchangesbecomepermanentinthesamewayas
theywouldifyouhadexecutedtheCOMMITstatement.
Here’sanexample:
Inthisexample,thechangeperformedwiththeUPDATEstatementhas
becomepermanent.Why?BecauseALTERTABLEisaDDLstatementand
carrieswithitanimplicitcommit.AsfarastheSHIPStableisconcerned,the
followingwouldhaveanequivalentimpact:
Inbothexamples,theUPDATEstatementiscommitted.Thefirstexample
demonstratesanimplicitcommit.Thesecondexampledisplayedpreviously
showsanexplicitcommit.FromtheperspectiveoftheSHIPStable,theultimate
effectisthesame—anALTERTABLEcommandandaCOMMITcommand
bothresultinacommiteventonthechangesperformedbytheUPDATE
statement,aswellasanypreviouslyperformedchangesthatmayhavenotyet
beencommittedsincethemostrecentofeitherasessionstartoranother
COMMITevent.
COMMITandMultipleSessions
Uptonow,we’velookedatCOMMITfromtheperspectiveofasinglesession.
Nowlet’sexpandourperspectiveanddiscussCOMMITanditsbehaviorwith
regardtomultiplesessionsloggedinsimultaneouslyandaccessingthesameset
oftables.Inshort,asingleuser,loggedintoasinglesession,canperform
changestoadatabaseusingINSERT,UPDATE,DELETE,oranycombination
ofthethree,butwhiletheresultingchangesarevisibleinthesessionwithin
whichthechangeswereperformed,theyarenotvisibletoanyotherlogin
sessionuntilaCOMMITisperformed.Thisincludesadditionalloginsessions
bythesameuserwhoperformedthechanges.Asingleusercanloginmultiple
timesandopenmultipleloginsessions,butuncommittedchangestodataare
visibleonlywithintheonesessionthatperformedthechanges,untilacommitis
performed—eitherexplicitlyorimplicitly.
Let’ssaythatoneuseraccountownsaschemathatincludesseveraltables,
thatallusershavebeengrantedprivilegestosee(SELECT)datafromthese
tables,andthatonlytheowninguseraccountretainstheprivilegeofmodifying
anydatainthosetables.IfthatowninguserexecutesaseriesofDMLstatements
withinasinglesession,anyresultingchangeswillbevisiblewithinthatsession
butwillnotbevisiblefromwithinanyothersessions—thatis,notuntila
commiteventoccursintheoriginalsessionwherethechangesweremade.Until
then,othersessionswillnotseethechanges.Theowningusermayperforma
largeseriesofINSERT,UPDATE,andDELETEstatementsandevenprint
reportsandscreenshotdisplaysshowingtheresultsofthemodifications,
performingSELECTstatementstodoit.However,unlessanduntilacommit
eventoccurswithinthesessionwherethechangeswereperformed,noother
session,regardlessofwhateverprivilegesmightbeinvolved,willbecapableof
seeingthechangeddata.Onlyaftersomesortofcommiteventoccurs—either
explicitorimplicit—dothechangesbecome“permanent,”asOraclelikestosay,
andthereforevisibletoallsessions.
Let’slookatavariationofthisscenario.LookatFigure3-7.Inthisscenario,
USER_1ownsatableSHIPSandhasgrantedUPDATEprivilegestoUSER_2.
ThismeansthatUSER_2,whodoesnotowntheSHIPStable,hastherightto
issueUPDATEstatementsandchangethatdata.So,USER_2issuesan
UPDATEstatementonSHIPSbutdoesnotcommitthechanges.Theresultis
thatifUSER_3hasSELECTprivilegesandtriestoquerytheSHIPStable
ownedbyUSER_1,thechangesmadebyUSER_2arenotvisibletoanyone
otherthanUSER_2.USER_3,forexample,cannotseethechangeddata.Inthis
example,thevalueforCAPTAIN_IDis0,andUSER_2issuesachangetothat
valueandupdatesitto7.ButUSER_3doesnotyetseethechange,sinceit
hasn’tbeencommittedtothedatabase.Forthatmatter,nouserseesthechange
—otherthanUSER_2.Noteventhetableowner,USER_1,willseeUSER_2’s
changeunlessanduntilUSER_2causesacommiteventtooccur.
FIGURE3-7 Uncommittedchange
Next,lookatFigure3-8.Here,USER_2hasissuedaCOMMITstatementto
createanexplicitcommitevent.TheresultisthatwhenUSER_3queriesthe
SHIPStable,thechangeissuedbyUSER_2isvisible.
FIGURE3-8 Committedchange
Inthisway,changesmadepriortoanycommiteventareinasortofstaging
area,wheretheusercanworkinwhatisalmosta“draft”mode.However,any
commitevent—explicitorimplicit—willmakechangespermanentandexpose
thenewdatatoallsessions.
ROLLBACK
TheROLLBACKstatementiscomparabletothe“undo”functioncommonto
manysoftwareapplications.ROLLBACKundoeschangestothedatabasethat
havebeenperformedwithinagivensession.Itdoesnotremoveanychangesthat
havealreadybeencommitted.Theonlychangesthatarerolledbackarethose
changesissuedwithinthesessionperformingtherollback.
Here’sanexample:
Inthisexample,oneINSERTstatementandoneDELETEstatementareissued.
Theresults:weaddonerowtothePORTStableanddeletealltherowsinthe
SHIPStable.
ButthenweissueaROLLBACKstatement,andbothchangesareeliminated.
ItisasifthosetwoDMLstatementsneverhappened.ThePORTSandSHIPS
tablesarerestoredtotheiroriginalconditionatthetimeofthelastCOMMIT
event.SincenoneoftheDMLstatementswascommitted,noothersessionsaw
thechanges.Thechangeswereobserved,albeittemporarily,onlybythesession
thatissuedthestatements.
ThechangesperformedbyuncommittedDMLstatementsarevisiblewithin
thesessionuntilthosechangesarerolledback.Forexample,seeFigure3-9.
FIGURE3-9 SamplesessionwithROLLBACK
Thefigureshowsthefollowingseriesofsteps:
AnexplicitCOMMITstatement
ASELECTtodemonstratethedatawithintheSHIPStable
AnUPDATEtochangedataintheSHIPStable
ThesameSELECTstatementtodemonstratethechangeddata
AROLLBACKtoremovetheeffectsoftheUPDATEstatement
ThesameSELECTstatementyetagain,showingthattheSHIPS
table’sconditionhasbeenrestored
Atnotimeduringthisprocessdidanyothersessionseetheeffectsofthe
UPDATEstatement.Thechangeswerevisibleonlytothesessionissuingthe
statements.
Ifaprogramabnormallyterminates,thedatabasewillissueanimplicit
rollback.Uncommittedchangesatthetimeofanabnormalterminationof,for
example,SQL*PlusorSQLDeveloperwillnotbecommittedtothedatabase.
SAVEPOINT
SofarwehavelookedattheTCLstatementsROLLBACKandCOMMIT.Let’s
lookatathird:SAVEPOINT.TheSAVEPOINTstatementestablishes
demarcationpointswithinatransactiontoempoweranyfollowingCOMMITor
ROLLBACKstatementstosubdividethepointsatwhichdatamaylaterbe
optionallysavedorundone.
WithoutSAVEPOINT,theCOMMITandROLLBACKstatementscan
operateonlyonasortofall-or-nothingbasis.Onceaseriesofstatementshave
beenexecuted,theentireseries,asonelargegroup,canbeeithersavedor
undone.ButifperiodicSAVEPOINTshavebeenissuedalongtheway,a
subsequentCOMMITorROLLBACKcanbeusedtosaveorrestoredatato
thosesavedpointsintimemarkedbyoneormoreSAVEPOINTstatements,thus
providingafinerlevelofdetailatwhichthetransactioncanbecontrolled.
HereisanexampleofSAVEPOINT:
Inthisexample,westartwithanexplicitCOMMITonline1andthenissuean
UPDATEstatement.Thenonline3,weissueaSAVEPOINTstatementand
nameitSP_1.That’sfollowedbyasecondUPDATEstatement.Giventhatwe
electedtoissuetheSAVEPOINT,wehaveanoptiononline5thatwehaven’t
seenyet,andthatisto“undo”thepreviousUPDATEstatement(butonlythat
secondUPDATE,notthefirst).Weaccomplishthisbyrollingbacktothe
SAVEPOINT.ThenweCOMMITourchanges.
TheresultisthatthevalueintheSHIP_ID12rowforHOME_PORT_IDis
21.
Here’sanotherexample;seewhetheryoucandeterminewhattheresulting
valueforHOME_PORT_IDwillbehere:
InthisseriesofSQLstatements,whatistheresultingvalueoftheSHIPS
table’sHOME_PORT_IDcolumnwheretheSHIP_IDvalueisequalto12?The
answer:22.ThatisthevaluecommittedtothedatabaseaftertheCOMMIT
statementonline8isexecuted.
Inthisexample,wecreatedtwoSAVEPOINTs—onewechosetoname
MARK_01,andanotherwechosetonameMARK_02.Wecouldhavechosen
anynameforthesesavepointsthatwewanted,accordingtotherulesofnaming
databaseobjects.Bynamingthem,wereservetherighttoselectivelyrollbackto
oneortheothernamedSAVEPOINT.Inthiscase,wechosetoissuea
ROLLBACKtotheSAVEPOINTnamedMARK_02.Thisstatementeffectively
restorestheconditionoftheentiredatabasetothepointwheretheSAVEPOINT
wasexecuted,which,inthisexample,ispriortotheUPDATEstatementonline
6.Inotherwords,it’sasiftheline6UPDATEstatementhadneverbeen
executed.
TherulesforusingSAVEPOINTincludethefollowing:
AllSAVEPOINTstatementsmustincludeaname.Thenamemust
bespecifiedusingthesamerulesandlimitationsfornamingdatabase
objects.Behindthescenes,theSAVEPOINTnameyoucreateis
associatedwithasystemchangenumber(SCN).Thisiswhatthe
SAVEPOINTismarking.
YoushouldnotduplicateSAVEPOINTnameswithinasingle
transaction—andrememberthatatransactionisaseriesofoneormore
SQLstatementsendingwithacommitevent.Ifyouduplicateaname,
youwillnotreceiveasyntaxorexecutionerror.Instead,thenew
SAVEPOINTwillsimplyoverridetheearlierSAVEPOINT,effectively
erasingit.
Onceacommiteventoccurs(eitheranexplicitorimplicitcommit
event),allexistingsavepointsareerasedfrommemory.Anyreferencesto
thembyfutureTCLstatementswillproduceanerrorcode.
Regardingthatlastpoint,here’sanexampleofwhatwe’retalkingabout:
Intheprecedingexample,theROLLBACKstatementonline5iswrongfor
tworeasons.First,it’slogicallyirrelevantsincethereisnothingtorollback—the
COMMITonline4permanentlysavedeverything,andnoadditionalSQL
statementshavebeenexecutedbythetimetheROLLBACKisexecutedonline
5.Second,theROLLBACKmakesareferencetoaSAVEPOINTthatdoesnot
exist,soitwillproduceanerror.Withoutthenamedsavepointreference,the
ROLLBACKwouldexecutejustfineandsimplyhavenoimpactonthe
database.ButwiththereferencetoanamedSAVEPOINTthatnolongerexists
bythetimetheROLLBACKisexecuted,theresultofline5isanerrorcode.
SAVEPOINTisparticularlyusefulwhenmanagingalargeseriesof
transactionsinwhichincrementalvalidationmustberequired,whereineach
individualvalidationrequiresacomplexseriesofDMLstatementsthatmight
failbutcanbeprogrammaticallycorrectedandvalidatedbeforemovingontothe
nextincrement.Agreatexampleofthisoccursinthefinancialworld.When
reconcilingasetoffinancialbooksforagivenorganization,youmightfindit
necessarytovalidateanaccountwithinalargerchartofaccounts.Duringeach
individualaccountvalidationprocess,theremaybeaneedtorollbacksome
statementsbeforemakingsomechangesandattemptingvalidationagain.Then,
onceagivenaccountisvalidated,youmightwanttodeclarethataccount
“temporarilyauthorized”andthenmoveontoattemptingtovalidatethenext
account.Furthermore,youmaywanttodeferacompleteCOMMITuntilall
accountshavebeenvalidated.SAVEPOINTisperfectforthis;aseachindividual
accountisvalidated,aSAVEPOINTcanbeestablishedsothatsubsequent
accountvalidationattemptsthatmightfailcantriggeraROLLBACKwithout
undoingtheearlieraccountsthatwerealreadysuccessfullyvalidated.Then,
whenallaccountsarevalidated—andnotbefore—asingleCOMMITcan
declarethefullsetofvalidatedrowstothedatabasewithapermanentsave.
ROLLBACKRevisited
Nowthatwe’veseenSAVEPOINT,let’stakeanotherlookatthesyntaxfor
ROLLBACKandseehowitcanbeusedtorollbacktoaparticularnamed
SAVEPOINT.
Sofar,we’veseenthisformoftheROLLBACKstatement:
ROLLBACK;
IfthesinglewordROLLBACKisexecutedbyitself,itwillignoreany
SAVEPOINTstatementsthatmayhavebeenexecutedsincethemostrecent
commiteventandundoanychangesmadebytheuserinthesessionsincethe
timeofthelastcommitevent.
However,ifROLLBACKisintendedtoundoanychangessinceanamed
SCNwasestablishedbySAVEPOINT,thenitcannametheSCNspecifically.
Forexample,ifaSAVEPOINTstatementestablishedademarcationwithan
SCNnamedscn_01,likethis:
SAVEPOINTscn_01;
thenasubsequentROLLBACKcanselectivelyundochangestothispoint
likethis:
ROLLBACKTOscn_01;
oralternativelylikethis:
ROLLBACKWORKTOscn_01;
Let’slookatthecomponentsofthisformoftheROLLBACKstatement:
First,thereservedwordROLLBACK
TheoptionalreservedwordWORK
TherequiredreservedwordTO
ThenameofanSCNasnamedbyaSAVEPOINTstatementthatwas
executedafterthemostrecentcommitevent
NotethattheWORKreservedwordisoptional.WORKispartoftheANSI
standardbutisnotrequiredintheOracleimplementation.
IfaROLLBACKstatementisexecutedthatnamesanonexistent
SAVEPOINT,SQLwilldisplayanerrorcodewarningthattherollbackwas
attemptingtorollbacktoaSAVEPOINTthatwasneverestablished.The
ROLLBACKwillfail,andnothingwillchangeregardingthestateofthe
database.Atthatpoint,anyoutstandingchangeswillremaininanuncommitted
state.Atthispoint,uncommittedchangesmaystillbecommittedorrolledback.
CERTIFICATIONSUMMARY
TheTRUNCATETABLEstatementcanbeusedinthosespecialcircumstances
whereyoumaywanttoremovealltherowsofatablewithoutfiringanyDML
triggersorselectivelyremovingdatafromanyassociatedINDEXobjects,as
wouldhappenwithaDELETEstatement.Insuchasituation,TRUNCATE
TABLEisanattractivealternativetodroppingandre-creatingatable,sinceit
leavesthetableanditsrelationshipsintact,aswellasanyindexes.TRUNCATE
TABLEcanremovestatementsmuchmorequicklywithlessprocessing
overheadthanDELETE.BecauseTRUNCATETABLEfiresnoDMLtriggers,it
isconsideredDDL;therefore,itsuseperformsanimplicitcommit,anditsresults
cannotberolledback.
IfTRUNCATETABLEisappliedtoatableforwhichthereexistsan
associatedparenttablewithaforeignkeyusingtheONCASCADEDELETE
clause,thenTRUNCATETABLEwillresultinanerrorifthereareinfactrows
inthatparenttablethatmightotherwisebedeletedasaresultofONCASCADE
DELETE.Forthissituation,youmustuseTRUNCATETABLE...CASCADE
toconfirmyourdesiretoremovethoseparentrows.
TheINSERTstatementaddsrowsofdatatoatable.Initssimplestform,it
addsonerowatatime.ItssyntaxstartswiththereservedwordsINSERTINTO
andthenameofasingledatabasetable,followedbyanoptionalcolumnlist,
followedbythereservedwordVALUES,followedbyalistofvaluestobe
insertedintothetable,enclosedinparentheses.Thelistofvaluesispresentedin
thesamesequentialorderasthecolumnlist,meaningthatfortherowthatthe
INSERTstatementisaddingtothetable,thefirstvalueinthelistwillbeadded
tothefirstcolumninthecolumnlist,thenthesecondvaluewillbeaddedtothe
secondcolumn,andsoon.Thedatatypesofthevaluesshouldbeappropriatefor
thedatatypesofthecolumnstowhichthevaluesarebeinginserted.Any
constraintsthatarenothonoredwillcausetheINSERTstatementtoberejected
atexecutiontime.Forexample,ifacolumnhasaNOTNULLconstraint,then
theINSERTstatementneedstoprovideavalidvalueforthatparticularcolumn,
orelsetheINSERTstatementwillfailandnorowisaddedtothetable.Ifthe
INSERTstatementomitsthecolumnlist,thenthevaluelistisassumedtobein
theorderofthecolumnsaccordingtothetargettable’sstructure,andeach
columnisassumedtobeaccountedforinthevalueslist.
TheUPDATEstatementisusedtomodifyrowsofdatainthetable.It
includestheoptionalWHEREclause,whichisusedtoidentifywhichrowsthe
UPDATEisintendedtomodify.IftheWHEREclauseisomitted,thenthe
UPDATEstatementwillattempttomodifyallrowsinthetargettable.The
UPDATEstatementusesthereservedwordSET,followedbypairsofcolumn
namesandvalues.Eachvaluecanbesubstitutedwithanexpression.Expressions
mayincludeliteralvalues,anyavailabletablecolumn,mathematicalequations,
andSQLfunctions.
TheDELETEstatementremovesrowsfromatable.TheoptionalWHERE
clausecanbeusedtoidentifywhichrowsaretobedeleted.However,ifthe
WHEREclauseisomitted,theneveryrowinthetablewillbedeleted.
TCL,whichisseparatefromDML,isoftenusedwithaseriesofDML
transactionstocontrolwhetherdataprocessedbythoseDMLtransactionswill
becommittedtothedatabase—inotherwords,made“permanent”asOracle
documentation(andothers)likestosay.TheTCLcommandsincludeCOMMIT,
ROLLBACK,andSAVEPOINT.
COMMITmakespermanentanyoutstandingchangestothetablesincethe
lastcommitevent.Commitscanoccurexplicitlyorimplicitly.Explicitcommits
occurwiththesimpleSQLstatementofCOMMIT.Implicitcommitsoccurwhen
othereventstakeplaceinthedatabase,suchasanyDDLstatement.AGRANT,
forexample,willautomaticallycommitallchangestothedatabasesincethelast
COMMIT.
ROLLBACKcanbeusedto“undo”aseriesofstatements.ROLLBACKused
byitselfundoesanychangesmadewithinasessionbytheusertothedatabase
sincethemostrecentcommitevent,implicitorexplicit,tookplace.Butifoneor
moreSAVEPOINTshavebeenissued,thentheROLLBACKmayoptionallyroll
backtoaparticularSAVEPOINT.
SAVEPOINTnamesasystemchangenumberandempowersfuture
executionsoftheROLLBACKstatementtogobacktoearlierversionsofthe
databaseincrementally.
✓TWO-MINUTEDRILL
TruncateData
TRUNCATETABLEremovesallofatable’srowsandallthedata
initsindexes,withoutfiringanytriggers.
ATRUNCATETABLEstatementisaDDLstatement.Thismeans
itsuseresultsinanimplicitcommitstatement,anditsresultscannotbe
rolledback.
IfyouuseTRUNCATETABLEonachildtableofaparentthat
usestheONDELETECASCADEclauseinitsforeignkeyconstraint,
thenyouusetheCASCADEclausewithTRUNCATETABLE,orelse
theexistingorchildrowswithmatchingparentrowswillcausethe
TRUNCATETABLEstatementtofailwithanerror.
InsertRowsintoaTable
TheINSERTstatementaddsoneormorerowstoatable.
TheINSERTsyntaxwereviewedinthischapterconsistsofthe
reservedwordsINSERTINTO,thenameofthetable,theoptional
columnlist,thereservedwordVALUES,andthelistofvaluestobe
entered.
IftheINSERTstatementiswrittensothatthelistofcolumnsinthe
tableisomitted,thenthelistofvaluesmustspecifyonevalueforeach
columninthetable;thosevalueswillbeprocessedinorderaccordingto
thecolumnsinthetable’sstructure.
ThelistofvaluesintheINSERTstatementmayinclude
expressions.Eachexpressionmustevaluatetoadatatypethatis
compatiblewithitstargetcolumninthetable.
Ifanyvalueviolatesanyconstraintappliedtothetargettable,then
anexecutionerrorwillresult.Forexample,eachnewrowaddedtothe
tablemustprovideanappropriatevalueofacompatibledatatypefor
eachNOTNULLcolumn;otherwise,anexecutionerrorisissued,and
theINSERTfails.
UpdateRowsinaTable
ThesingleUPDATEstatementcanmodifyexistingdatainoneor
morerowswithinadatabasetable.
TheUPDATEstatementsyntaxstartswiththereservedword
UPDATEandthenameofthetargettable,thereservedwordSET,and
thenaseriesofassignmentexpressionsinwhichtheleft-sideelement
specifiesatablecolumn,followedbytheassignmentoperator(anequal
sign),followedbyanexpressionthatevaluatestoadatatypeappropriate
forthetargettable’scolumnidentifiedontheleftsideoftheequalsign,
andfinallyanoptionalWHEREclause.
Ifadditionalassignmentexpressionsarerequired,eachadditional
assignmentexpressionisprecededbyacomma.
IftheWHEREclauseisomitted,thenalltherowsinthetableare
changedaccordingtotheseriesofSETvalueslistedintheUPDATE
statement.
DeleteRowsfromaTable
TheDELETEstatementisusedtoremoverowsofdatafroma
table.
ThesyntaxstartswiththereservedwordsDELETEandtheoptional
FROM,thenthenameofthetargettable,thenanoptionalWHERE
clause.
IftheWHEREclauseisomitted,alltherowsinthetableare
deleted.
ControlTransactions
TCLstatementsincludeCOMMIT,ROLLBACK,and
SAVEPOINT.
Therearetwotypesofcommitevents:explicitcommitandimplicit
commit.
AnexplicitcommitoccurswiththeCOMMITstatement.
Animplicitcommitoccursimmediatelybeforeandaftercertain
eventsthattakeplaceinthedatabase,suchastheexecutionofanyvalid
DDLstatement,includingCREATE,ALTER,DROP,GRANT,and
REVOKE.EachDDLstatementexecutionisautomaticallyprecededand
followedbyanimplicitcommit.
TheCOMMITstatementisusedtosavechangesperformedwithin
thesessionandmakethosechanges“permanent.”
IfaDDLstatementfailsduringexecution,theimplicitcommitthat
precededitstillisineffect,ensuringthatthecommitoccurred,whether
theDDLstatementwassuccessfulornot.ThesameisnottrueforDDL
statementsyntaxerrors.
TheROLLBACKstatementisusedtoundochangestothe
database.
TheSAVEPOINTstatementcanbeusedtonameapointwithina
seriesofSQLstatementstowhichyoumayoptionallyrollbackchanges
afteradditionalDMLstatementsareexecuted.
OnceaCOMMITisissued,allexistingSAVEPOINTsareerased.
AnyROLLBACKthatnamesnonexistingSAVEPOINTswillnot
execute.
IfROLLBACKisissuedwithoutnamingaSAVEPOINT,changes
madebytheuserduringthecurrentsessionarerolledbackwithinthat
sessiontothemostrecentcommitevent.
SELFTEST
Thefollowingquestionswillhelpyoumeasureyourunderstandingofthe
materialpresentedinthischapter.Choosethebestsingleanswerforeach
questionunlessotherwisespecified.
TruncateData
1.ReviewthefollowingSQLstatement:
TRUNCATEpersonnel;
Whichofthefollowingistrueofthepreviousstatement?(Chooseall
thatapply.)
A.Thestatementwillresultinanimplicitcommit.
B.ThestatementwillremovealldatafromanyINDEXobjects
associatedwiththattable.
C.ThestatementwillnotfireanyDMLtriggersonthetable.
D.Thestatementwillfail.
2.TheCASCADEkeyword,whenusedwithTRUNCATE:
A.Isrequiredifthetablehasanydependentchildtables
B.Willensurethatfutureattemptstoinsertrowstothetablewillbe
rejectediftheysatisfytheTRUNCATEtable’sWHEREclause
C.CanbeusedwiththeoptionalDEPENDENCYkeyword
D.Noneoftheabove
3.TRUNCATETABLE:
A.CannotbeusedwithinavalidSQLstatement
B.IsavalidsetofkeywordstobeusedwithinaDDLstatement
C.DoesnotrequiretheDROP_ANY_TABLEprivilege
D.IsavalidstatementthatwilltruncateatablecalledTABLE
InsertRowsintoaTable
4.Reviewthefollowingstatement:
Thetablewillcreatesuccessfully.WhatwillresultfromtheINSERT
statement?
A.TheINSERTwillfailbecausethereisnolistofcolumnsafter
STUDENT_LIST.
B.TheINSERTwillfailbecausetheliteralvalueforPHONEis
numericandPHONEisacharacterdatatype.
C.TheINSERTwillexecute—thetablewillcontainonerowof
data.
D.Noneoftheabove.
5.ConsiderthefollowingsetofSQLstatements:
Thetablewillcreatesuccessfully.Whatwillbetheresultofthetwo
INSERTstatements?
A.Neitherwillexecute.
B.Thefirstwillexecute,butthesecondwillfail.
C.Thefirstwillfail,butthesecondwillexecute.
D.Bothwillexecutesuccessfully.
6.ConsiderthefollowingsetofSQLstatements:
WhatwillbetheresultoftheINSERTstatement?
A.ItwillfailbecausethereisnocolumnlistintheINSERT
statement.
B.ItwillfailbecausethereisnoPRIMARYKEYinthetable.
C.Itwillexecuteandcreateanewrowinthetable.
D.Itwillfailbecausethelastnameandfirstnamevaluesare
reversed.
UpdateRowsinaTable
7.Whichofthefollowingreservedwordsisnotrequiredinordertoform
asyntacticallycorrectUPDATEstatement?
A.UPDATE
B.SET
C.WHERE
D.Noneoftheabove
8.AssumeatableLAMPSthathasnoconstraints.Whichofthe
followingistrueabouttheUPDATEstatementandtheLAMPStable?
(Chooseallthatapply.)
A.UPDATEcanbeusedtoaddrowstoLAMPSbysettingvalues
toallthecolumns.
B.UPDATEcanbeusedtoremovearowfromLAMPSbysetting
alloftherow’scolumnstoavalueofNULL.
C.ForexistingrowsinLAMPS,UPDATEcanaddvaluestoany
columnwithaNULLvalue.
D.ForexistingrowsinLAMPS,UPDATEcanremovevaluesfrom
anycolumnbychangingitsvaluetoNULL.
9.ReviewthefollowingSQLstatements:
Whatcanbesaidofthestatementslistedhere?
A.Onerowwillbeupdated.
B.Tworowswillbeupdated.
C.Atleastoneofthestatementswillnotexecute.
D.Noneoftheabove.
10.ReviewthefollowingSQLstatements:
WhatistheendresultoftheSQLstatementslistedhere?
A.TheBOUNCERStablewillcontainonerow.
B.TheBOUNCERStablewillcontaintworows.
C.TheUPDATEwillfailbecausethereisnoWHEREclause.
D.Noneoftheabove.
DeleteRowsfromaTable
11.WhichofthefollowingreservedwordsisrequiredinacompleteDELETE
statement?(Chooseallthatapply.)
A.FROM
B.WHERE
C.DELETE
D.Noneoftheabove
12.ConsiderthefollowingdatainatablecalledPARTS:
WhichofthefollowingSQLstatementswillremovethewordVALID
fromrow1,resultinginonerowwithastatusofNULLandtworowswith
astatusofPENDING?
13.ReviewthefollowingSQLstatements:
Whichofthefollowingbestdescribestheresultsofattemptingto
executetheDELETEstatement?
A.TheDELETEstatementwillfailbecauseitismissingacolumn
listbetweenthewordDELETEandthenameofthetable
AB_INVOICES.
B.TheDELETEstatementwillexecute,butnorowsinthetable
willberemoved.
C.TheDELETEstatementwillproduceasyntaxerrorbecauseitis
referencingarowthatdoesnotexistinthedatabase.
D.Noneoftheabove.
ControlTransactions
14.Assumeaschemawithonlytwotables:onenamedPRODUCTSandone
namedENGINEERING.ReviewthefollowingSQLstatements:
InthisseriesofSQLstatements,whichlinerepresentsthefirstcommit
event?
A.Line1
B.Line2
C.Line4
D.Line6
15.ReviewtheSQLstatementsthatfollow,andassumethatthereisnotable
calledADDRESSESalreadypresentinthedatabase:
WhatwillbetheresultoftheexecutionoftheSQLstatementsshown
here?
A.TheADDRESSEStablewillhaveonerowwithavalueof1for
ZONE.
B.TheADDRESSEStablewillhaveonerowwithavalueof2for
ZONE.
C.TheADDRESSEStablewillhavenorows.
D.Noneoftheabove.
SELFTESTANSWERS
TruncateData
1. D.Thestatementwillfail.TRUNCATETABLEiswhatisintended
here.TRUNCATEbyitselfisnotavalidstatement.
A,B,andCareincorrect.Theywouldbecorrectifthestatement
includedthekeywordTABLE.Butwithoutthatkeyword,thestatementwill
fail,sotheseoptionsarenotapplicable.
2. D.ThepurposeofCASCADEistoensurethatanydependentchild
tableswithrowsthatrelatetothetruncatedrowswillalsobetruncated.
A,B,andCareincorrect.CASCADEisnotrequired;itisoptional.
TRUNCATEhasnoeffectonsubsequentSQLstatements,including
INSERT.ThereisnoDEPENDENCYkeywordusedwithTRUNCATE.
3. B.Thesekeywords,whencombinedwithavalidtablename,will
formavalidDDLstatement.
A,C,andDareincorrect.Thesearevalidkeywords.TRUNCATE
requirestheDROP_ANY_TABLEprivilege.TABLEisakeywordand
requiredforTRUNCATE,anditrequiresavalidtablenametofollowthe
keywordTABLE.
InsertRowsintoaTable
4. C.Thestatementsaresyntacticallyandlogicallycorrect.The
INSERTstatementomitsthecolumnlist.Thisomissionrequiresthelistof
valuestobeprovidedinthesamesequenceinwhichthecolumnsappearin
thetable’sstructure,asindicatedintheCREATETABLEstatement.
A,B,andDareincorrect.ThevalueprovidedbyINSERTforthe
PHONEcolumnisanexpressionthatiscompatible;theVARCHAR2data
typeisacharacterdatatype,anditacceptsbothnumericandtextdata.In
fact,valueslikephonenumbersandZIPcodesarebesttreatedascharacter
data;otherwise,leadingzeroswillbetruncated,andinthecaseofaZIP
code,thatcanbeadisasterforalotofaddressesintheNortheasternUnited
States.
5. D.Thesyntaxisfine,andbothINSERTstatementswillexecute.
A,B,andCareincorrect.
6. C.Itwillcreateanewrowinthetable.Thefactthatthecolumn
valuesareprobablyreversedmayrepresentalogicalerrornowandcreate
problemsdowntheroad,butthere’snothingaboutthestatementthatwill
preventitfromexecutingsuccessfully.
A,B,andDareincorrect.ThelackofacolumnlistintheINSERT
merelyrequirestheretobealistofvaluesmatchingthenumberanddata
typesofthecolumnsinthetable’sstructure,andthisINSERTstatement’s
listofvaluessatisfiesthatrequirement,albeitinanapparentlyillogicalway,
butnevertheless,therequirementsforSQLaremet.Thelackofa
PRIMARYKEYonthetableprobablyrepresentspoordesign,butisnota
problemwithregardtothesuccessfulexecutionoftheSQLstatementshere.
UpdateRowsinaTable
7. C.AnUPDATEstatementdoesnotneedtohaveaWHEREclause.
IfaWHEREclauseisomitted,theneveryrowinthetargettableissubject
tobechangedbytheUPDATEstatement,dependingonwhetherany
constraintsexistonthetableandwhethertheypermitorrejectthedata
changestheUPDATEstatementisattemptingtomake.
A,B,andDareincorrect.ThereservedwordUPDATEisrequiredfora
validUPDATEstatement.ThesameistrueforthereservedwordSET.
8. CandD.Addingavaluetoacolumninanexistingrowisthe
purposeoftheUPDATEstatement.SettingavaluetoNULLisas
acceptableassettingavaluetosomeotherspecificvalue.
AandBareincorrect.INSERTaddsnewrowstoatable,andDELETE
removesthem.UPDATEdoesn’tremoverowsfromatable.TheUPDATE
statementcanmodifyonlyexistingrows.InthequestionoftheLAMPS
table,ifyouchoosetouseSETtoseteachcolumn’svaluetoNULL,then
you’llstillhavearowinthetable;itwillsimplyconsistofNULLvalues.
Butyou’llstillhavearow.AndyoucannotcreateanewrowbyusingSET
tosetvaluestoeachcolumn;allyoucandoismodifyexistingrows.And,
ofcourse,ifLAMPSwerecreatedwithanyNOTNULLconstraints,any
UPDATEstatementwouldhavetorespectthoseconstraints.Butthe
questionassertedtheLAMPShasnoconstraints.
9. C.TheUPDATEstatementcontainsanextraoccurrenceofthe
reservedwordSET.OnlythefirstSETbelongs;thesecondshouldbe
removed.Isthisatrickyquestion?Yes,itis.Soaremanyofthequestions
onthecertificationexam.Makesureyourknowledgeofsyntaxisstrong.
Andreadcarefully.
A,B,andDareincorrect.
10. A.TheINSERTstatemententersasinglerow,andtheUPDATEstatement
modifiesthatsinglerow,leavingonemodifiedrowinthetable.
B,C,andDareincorrect.Thereisonlyonerowinthetable—the
UPDATEdoesnotaddanewrowbutratherchangestheexistingrow.
UPDATEdoesnotrequireaWHEREclause;withoutit,theUPDATE
statementappliesitschangestoalltherowsinthetable,andthistable
containsonerow.
DeleteRowsfromaTable
11. C.TheonlyrequiredreservedwordisDELETE.
A,B,andDareincorrect.FROMisoptional.WHEREisalsooptional.
12. Discorrect.DELETEremovesentirerowsfromthedatabase.Removing
asinglevaluefromasinglecolumnrequirestheuseoftheUPDATE
statement.
A,B,andCareincorrect.AandBarevalidDELETEstatements,either
ofwhichwillremovethefirstrowfromthetable,insteadofjustremoving
thevalueforthe“status”column.OptionCisaninvalidstatementthatwill
triggerasyntaxerror—theSETreservedwordhasnoplaceintheDELETE
statement.
13. B.Thesyntaxisfine,andthestatementwillexecuteasintended,whichis
toremoveanyrowsfromthetablewithanINVOICE_IDvalueof2.Itjust
sohappensthattherearen’tanyrowsthatmatchthestatedcriteriaatthe
timetheDELETEstatementisissued.
A,C,andDareincorrect.ThereisnocolumnlistinaDELETE
statementbeforethetablename.AndthefactthattheWHEREclausedoes
notidentifyanyrelevantrowsisnotasyntaxproblem,norisita
compilationproblem—thestatementwillsimplynotdeleteanyrows.
ControlTransactions
14. B.Line2isaDROPstatement,whichfallsunderthetypeofSQL
statementsknownasDataDefinitionLanguage.AllDDLstatementscause
animplicitcommittooccur,eveniftheDDLstatementfailsinexecution.
A,C,andDareincorrect.TheSELECTstatementhasnoimpactona
commiteventatall.Line4isanexplicitCOMMIT,andwereitnotforline
2,thiswouldbethefirstcommiteventinthissetofstatements.Line6
undoestheeffectsofline5andundoestheuser’schangestothedatabase
sincethepreviouscommitevent,whichatthisstageisrepresentedbythe
line4commit.
15. C.TheROLLBACKstatementdoesnotreferencetheSAVEPOINTname,
soinsteaditrollsallthewaybacktothelastCOMMITevent,whichinthis
caseistheimplicitcommitthatoccurredwiththeCREATETABLE
statement.
A,B,andDareincorrect.
4
RestrictingandSortingData
T
CERTIFICATIONOBJECTIVES
4.01SorttheRowsThatAreRetrievedbyaQuery
4.02LimittheRowsThatAreRetrievedbyaQuery
4.03UseAmpersandSubstitutiontoRestrictandSortOutputatRunTime
4.04UsetheSQLRowLimitingClause
✓Two-MinuteDrill
Q&ASelfTest
hischapterlooksatvariouscapabilitiesoftheSELECTstatement:the
WHEREclause,theORDERBYclause,ampersandsubstitution,andthe
SQLrowlimitingclause.TheORDERBYclausesortstheretrievedrows.It’s
veryflexible:itcansortrowsinascendingordescendingorder,oritcansortby
expressionlistsortakeadvantageofotherpowerfulfeatures.TheWHERE
clausespecifiesthecriteriarequiredforarowtobeincludedinaSQLstatement.
Withoutit,allrowsinagiventableareretrieved,butwithit,aSQLstatement
canselectivelytargetparticularrowsforprocessing.Ampersandsubstitution
providesuniqueandpowerfulcapabilitieswhenworkingwithOracleSQL.The
SQLrowlimitingclauseisaflexibleenhancementtotheSELECTstatement.
AmpersandsubstitutionisaSQL*Pluscapabilityontheexam.Weknowitisa
SQL*Plusstatementbecauseitisnotdescribedinthe"OracleDatabaseSQL
LanguageReference,"butisdetailedinsteadinthe“SQL*PlusUser'sGuide
andReference."Ampersandsubstitutionprovidesuniqueandpowerful
capabilitieswhenworkingwithOracle’sSQL,thusitsinclusionontheexam.
Aworkingknowledgeoftheseclausesisnecessarytopasstheexam.Let’s
getstarted.
CERTIFICATIONOBJECTIVE4.01
SorttheRowsThatAreRetrievedbyaQuery
ThissectionlooksatanotherclauseintheSELECTstatement,theORDERBY
clause.ORDERBYisusedtosorttherowsthatareretrievedbyaSELECT
statement.Itsortsbyspecifyingexpressionsforeachrowinthetable.Sorting
canbeperformedineitherascendingordescendingorder.SQLwillsort
accordingtothedatatypeoftheexpressionthatisidentifiedintheORDERBY.
Youcanincludemorethanoneexpression.Thefirstexpressionisgivensorting
priority,thesecondisgivensecond-positionpriority,andsoon.
ORDERBYisalwaysthefinalclauseinaSELECTstatement.Itisusedonly
inSELECT;contrarytotheWHEREclause,whichcanalsobeusedinUPDATE
andDELETE,theORDERBYclauseisuniquetotheSELECTstatementandis
notusedintheotherSQLstatements.
(NotethatinChapter9we’llexaminehowyoumayembedaSELECT
statementasasubquerywithinanINSERT,UPDATE,orDELETEstatement—
sointhatregard,itistheoreticallypossiblethatanORDERBYclausemightbe
includedinaSELECTstatementthatisembeddedwithin,forexample,an
INSERTstatement.Butthatisaseparateissue.)
ORDERBYdoesnotchangedataasitisstoredinthetable.Datainatable
remainsunchangedasaresultoftheORDERBY.ORDERBYispartofthe
SELECTstatement,andtheSELECTstatementisincapableofchangingdatain
thedatabase.Note,however,thatwhenSELECTisembeddedwithinotherSQL
statementslikeINSERTorUPDATE,changestothedatabasecanresult,butnot
becauseofSELECTalone.You’llseehowthatworksinChapter9.
ORDERBYsortstheoutputofaSELECTstatementfordisplaypurposes
only.ItisalwaysthelaststepinaSELECTstatementandperformsitssortafter
allthedatahasbeenretrievedandprocessedbytheSELECTstatement.
ReferencebyName
Let’slookatanexampleofORDERBYinaction.Considerthefollowingdata
listingfromatableADDRESSES:
Wecanselectthisdataandsortitbyspecifyingthecolumnname(ornames)
intheORDERBYclauseoftheSQLstatement,asfollows:
Figure4-1showstheresults.Noticeinthefigurethattherowsaresortedin
alphabeticalorderaccordingtothevalueintheSTATEcolumn.
FIGURE4-1 ResultsofSELECTwithORDERBYSTATE
NotethattherowwithaNULLvalueforSTATEislast.TheNULLvalueis
consideredthe“highest”value.
Aswereviewtheoutput,itbecomesclearthatwemightalsowantto
alphabetizeourinformationbycityforeachstate.Wecandothatbymodifying
ourORDERBYclauseeversoslightly:
InthemodifiedversionofourORDERBY,weaddasecondcolumntothe
ORDERBYclausebywhichwewanttosort.Figure4-2displaystheresultsof
thisSELECTstatement.TherowshavebeensortedfirstbySTATEandthenby
theCITYvalueforeachSTATE.
FIGURE4-2 ResultsofSELECTwithORDERBYSTATE,CITY
NotethatthechoiceofcolumnsweincludeintheORDERBYclausedoes
notinfluencewhichcolumnswechoosetodisplayintheSELECTexpression
list.Itwouldprobablybeeasiertoreadifweputthesamecolumnsusedinthe
ORDERBYinthesamepositionsastheSELECTexpressionlist.Inother
words,thisquerywouldprobablyproduceamorereadableoutput:
TheoutputofthisSELECTwoulddrawattentiontoourintent,whichisto
sortdatabySTATEfirstandthenCITY.Butwhilethismightbeconsidered
preferentialdesignincertaincircumstances,itisbynomeansrequiredwithin
thesyntaxoftheSQLstatement.We’renotevenrequiredtoincludetheORDER
BYcolumnsintheSELECTstatement’sexpressionlistatall.Thisisanother
perfectlyvalidSELECTstatement:
Noticethatwe’resortingbycolumnsthataren’tincludedintheSELECT
statement’sexpressionlist.
WithoutanORDERBYclause,thereisnoguaranteeregardingthesequence
inwhichrowswillbedisplayedinaSELECTstatement’soutput.Therowsmay
beproducedinadifferentorderfromonequerytoanother.Theonlywayto
ensureconsistencytotheoutputistoincludeanORDERBYclause.
ASCandDESC
Therearetworeservedwordsthatspecifythedirectionofsortingonagiven
columnoftheORDERBYclause.ThosereservedwordsareASCandDESC.
ASCisshortfor“ascending”andindicatesthatvalueswillbesorted
inascendingorder.Inotherwords,thelowest,orleast,valuewillbe
listedfirst,followedbyvaluesofhigher,orgreater,value.ASCisthe
defaultchoiceandassuchdoesnotneedtobespecifiedwhendesiredbut
maybespecifiedforclarity.
DESCisshortfor“descending”andindicatesthatvalueswillbe
sortedindescendingorder.Inotherwords,thehighest,orgreatest,value
willbelistedfirst,andvalueswillcontinuetobelistedindecreasing,or
lesser,value.
Asjuststated,thedefaultisASC.Youdon’tneedtospecifyASC.The
ORDERBYexamplesyou’veseensofarhavedefaultedtoASCwithoutour
havingtospecifyit.ButyoucanspecifyASCifyouwish.
Here’savariationontheprecedingSELECTstatementthatusesa
combinationofASCandDESC:
Figure4-3displaystheresultsofthisSELECT.NoticethattheSHIP_ID
valuesarelistedinascendingorder,butthatforeachship,thePROJECT_COST
valuesareshownfromthehighesttothelowestvalues.
FIGURE4-3 PROJECTStablewithORDERBYASCandDESC
ASCandDESCeachoperateontheindividualORDERBYexpressions.
ThereisnowaytoassignASCorDESCtoalltheORDERBYexpressions
collectively;instead,youmustplaceyourchoiceaftereachindividualORDER
BYexpression,rememberingthatASCisthedefaultandthereforedoesnotneed
tobespecified.
Expressions
NotethattheexpressionsyoucanincludeinanORDERBYclausearenot
limitedtocolumnsinatable.Anyexpressionmaybeused.Tobeuseful,the
expressionshouldincludeavalueinthetable;otherwise,thevaluemaynot
change,andtherewon’tbeanymeaningfuleffectontherowsinthetable.
Forexample,here’sadatalistingforthePROJECTStable:
Thisdatalistingshowsaseriesofprojects.Foreachproject,weseethe
SHIP_IDforwhichtheprojectisintended,theproject’stotalcost,andthe
estimatednumberofdaysitwilltaketocompleteeachproject.
LookingatthevaluesforPROJECT_COSTandDAYS,weseeenough
informationtocomputetheaveragecostperdayforeachproject.
Forexample,theper-daycostforthethree-day“LifeboatInspection”willturn
outtobe4000,or12000dividedby3.
Whatifwewanttosorttheserowsaccordingtothecomputedvalueofthe
PROJECT_COST/DAYS?Noproblem.
Thatquerywillachievetheresultwe’reafter.Let’svaryitabittoseemoreof
whatitiswe’recalculatingandsorting.
Figure4-4showstheresultsofthisquery.Notethattherowsareorderedbya
valuethatdoesn’texistinthetablebutratheravaluethatistheresultofan
expressionthatdrawsvaluesfromthetable.
FIGURE4-4 PROJECTSsortedbyPROJECT_COST/DAYS
TheColumnAlias
Asyou’vealreadyseen,aSELECTstatementcanincludeexpressionsinthe
selectlist.Here’sanexample:
NoticetheoutputinFigure4-4andthedefaulttitleofthefifthcolumn.SQLhas
usedtheexpressionasthetitleofthecolumn.
WecouldhaveusedaSQLfeaturecalledthecolumnalias.Hereisthesame
querywithacolumnalias(linenumbersadded):
NoticethePER_DAY_COSTcolumnaliasattheendofline2.Thecolumn
aliasisanameyoumakeupandplacejustafterthecolumnyouwanttoalias,
separatedbytheoptionalkeywordAS.Inthisexample,thecolumnwiththe
columnaliasisthefinalcolumnintheselectlist.Oncethecolumnaliasisused,
youcanreferenceitfromwithintheORDERBYclause,aswedoonline4.
Therulesforusingacolumnaliasincludethefollowing:
EachexpressionintheSELECTlistmayoptionallybefollowedbya
columnalias.
Acolumnaliasisplacedaftertheexpressionintheselectlist,
separatedbytheoptionalkeywordASandarequiredspace.
Ifthecolumnaliasisenclosedindoublequotes,itcaninclude
spacesandotherspecialcharacters.
Ifthecolumnaliasisnotenclosedindoublequotes,itisnamed
accordingtothestandardrulesfornamingdatabaseobjects.
ThecolumnaliasexistswithintheSQLstatementanditsresultsand
doesnotexistoutsideoftheSQLstatement.
Thecolumnaliasmustbeunique;itwillbecomethenewheaderin
theoutputoftheSQLstatement.
ThecolumnaliascanbereferencedwithintheORDERBYclause,
butnowhereelse—suchasWHERE,GROUPBY,orHAVING.
Here’sanexampleofacolumnaliasthatusesthedoublequotationmarks.
Noticetheinclusionofaspaceinthealias.
SeetheoutputofthisqueryinFigure4-5.Noticethecolumnheadingforthe
aliasedcolumn—thealiasbecomesthenewheadingintheSQLoutput.
FIGURE4-5 SELECToutputwithcolumnalias
ThepointofbringingupthealiashereinthisdiscussionaboutORDERBYis
this:youcanusethecolumnaliaswhenreferencinganycolumnwithORDER
BY,andit’sparticularlyusefulwhentryingtouseORDERBYwithan
expressionfromwithintheSELECTstatement’sexpressionlist.
ReferencebyPosition
AnotherwaytheORDERBYclausecanidentifycolumnstobesortedisviathe
“referencebyposition”method.ThisworksonlyiftheintentoftheORDERBY
istosortrowsaccordingtoinformationthatisincludedintheSELECTlist.
Here’sanexample:
NoticethatwechoosetoORDERBY5inthisSQLstatement.Thenumber5in
thiscontextisreferencingthefifthitemintheSELECTstatement’sselectlist,
whichistheexpressionPROJECT_COST/DAYS.
AnyexpressionintheSELECTlistcanbereferencedusingitsnumeric
position.Thefirstexpressionisconsiderednumber1,thesecondisnumber2,
andsoon.
Anyattempttoreferenceapositionthatdoesn’texistwillproduceaSQL
error;forexample,thisisinvalid:
Thisstatementwillnotexecute.TheORDERBYclausemustidentifyanumber
thatcorrespondstoanitemthatisintheSELECTlist.
Combinations
ORDERBYcancombinethevarioustechniquesofreferencebyname,reference
bycolumnalias,andreferencebyposition.Here’sanexample:
Thisexampleisavalidstatement.Itsortsrowsby
ThevalueofSHIP_ID,indescendingorder
ThevalueinthePROJECT_NAMEcolumn,whichhasacolumn
aliasinthisSQLstatementof"TheProject"
ThevalueinthePROJECT_COSTcolumn,whichistheseconditem
intheSELECTlist
Figure4-6showstheoutputofthisquery.
FIGURE4-6 PROJECTSsortedbymultipletechniques
RememberthatorderingaSELECTstatementbypositionisextremely
usefulinmanysituationsthatinvolvecomplexSELECTstatements.
Lateryou’llseesomeinterestingcombinationsofmultipleSELECT
statements(suchasthesectionthatlooksatusingsetoperatorsthat
combinemultipleSELECTstatementsintoone),andinthosesituations,
youcanalwaysreferenceanORDERBYcolumnbyposition.
ORDERBYandNULL
WhenSELECTperformsasortusingORDERBY,ittreatsanyvaluesthatit
mightfindtobeNULLas“greaterthan”anyothervalue.Inotherwords,when
yousortbyanumericdatatypecolumnandthatcolumncontainsNULLvalues,
theNULLvalueswillsortasbeinggreaterthanallNOTNULLvaluesinthelist.
Thesameistrueforcharacterdatatypesanddatedatatypes.
CERTIFICATIONOBJECTIVE4.02
LimittheRowsThatAreRetrievedbyaQuery
WhenbuildingaSELECTstatement,yourfirsttaskistoidentifywhichdatabase
table(ortablesorviews)containsthedatayouneed.Youalsoneedtolookatthe
table’sstructuretochoosethecolumnsthatwillbeincludedintheselectlistof
yourSELECTstatement.Butrarelydoyoustopthere.Mostofthetimeyou’ll
wanttolimittherowsyou’llretrievetoaparticularfew,basedonsomesortof
businessrules.ThattaskisaccomplishedwiththeSELECTstatement’sWHERE
clause.
ThissectionwilllookattheWHEREclauseanditsusage.
TheWHEREClause
TheWHEREclauseisoneofthemoreimportantclausesofthreedifferentSQL
statements:SELECT,UPDATE,andDELETE.
ThepurposeoftheWHEREclauseistoidentifyrowsthatyouwantto
includeinyourSQLstatement.Ifyou’reworkingwithaSELECT,thenyour
WHEREclausechooseswhichrowswillbeincludedinyourSELECToutput.If
it’sanUPDATEyou’reworkingwith,theWHEREclausedefineswhichrows
willbeupdated.Ifit’saDELETE,theWHEREclausedefineswhichrowswill
bedeleted.
WithinanySQLstatement,theWHEREclause,ifincluded,alwaysfollows
theFROMclause.WHEREisoptional;itisneverrequiredinordertoforma
completeSQLstatement,butifincluded,itmustfollowtheFROMclause.
TheWHEREclausestartswiththereservedwordWHERE,whichis
followedbytheWHEREcondition.TheWHEREconditionconsistsofoneor
morecomparisonsofexpressions.TheultimategoaloftheWHEREconditionis
todetermineavalueoftrueorfalseforeachrowinthetables(and/orviews)
identifiedintheFROMclause.IftheWHEREconditionreturnsatruefora
givenrow,thatrowisincludedintheSQLstatement.Ifitreturnsafalsefora
givenrow,thatrowisignoredfortheSQLstatement.
Let’slookatasimpleexample:
Inthisexample,theWHEREclauseonline3comparestwoexpressions:
ThefirstexpressionisthetablecolumnSHIP_ID.
Thesecondexpressionconsistsoftheliteralvalue3.
Thecomparisonoperatorusedtocomparethesetwoexpressionsistheequal
sign.TheWHEREclausewillconsidereachrowinthetables(and/orviews)
identifiedintheFROMclause.ForthisWHEREclause,eachrow’sSHIP_ID
valueisanalyzedtodeterminewhetheritsvalueequals3.Foreachrowthat
containsavalueof3intheSHIP_IDcolumn,theWHEREconditionreturnsa
valueof“true”forthatrow,andthatrowisincludedintheSQLstatement.All
otherrowsareignored.
NotethattheSELECTstatement’sselectlistdoesn’tincludeSHIP_ID.The
WHEREclausedoesnotneedtoincludeanyofthecolumnsthatarebeing
displayed—anycolumninthetableisavailabletotheWHEREclause.Inthis
particularexample,theissueofwhethertoincludearowisbasedondatathat
won’tbeincludedinthefinalresultsoftheSELECTstatement,andthat’sfine.
IfyouleavetheWHEREclauseoutofaSELECTstatement,thenallrowsof
thetable(ortables)areretrieved.
HereisanotherexampleofaWHEREclause:
ThisexampleshowsacompleteWHEREclauseonline3.Inthisexample,
theSELECTstatementwillshowthevaluesforPORT_NAMEandCAPACITY
forallrowswithavalueintheCAPACITYcolumnthatisgreaterthanorequal
to5.
Inthenextsectionwe’llexamineexpressionsinWHEREclauses.
ComparingExpressions
TheWHEREclauseusesaseriesofcomparisons,eachofwhichevaluatesfor
eachrowtoeithertrueorfalse.Eachcomparisoninvolvestwoexpressions
evaluatedbyacomparisonoperator.Theexpressionsarefirstevaluatedandthen
compared.Forexample,inthefollowingcomparison,ontheleftisanexpression
thatconsistssolelyofthecolumnnamedSALARY,andontherightisan
expressionthatmultipliestwonumberstogether:
Inthisexample,ifthevalueinSALARYisgreaterthanorequaltotheresultof
themathequation50899times1.12,theresultistrue.Otherwise,it’sfalse.
Theexamplesyou’veseensofarhaveshowncolumnnamesontheleft,but
that’snotrequired;anyvalidexpressionmaybeplacedoneitherside—orboth
sides—ofthecomparisonoperator.Ideallyoneoftheexpressionsshouldinclude
datafromthetable;otherwise,what’sthepoint?Butintermsofsyntax,thatis
notrequired.Allthatisrequiredisavalidexpressiononbothsidesofthe
comparisonoperator.
Let’slookatanotherexample:
Foragivenrow,ifthevalueinthecolumntitledSTART_DATEislessthanthe
valueinthecolumntitledEND_DATE,theexpressionistrue;otherwise,it’s
false.(NotethatwhenSQLcomparesdates,“lessthan”means“earlierthan.”
Forexample,January1islessthanJanuary2ofthesameyear.We’lltalkmore
aboutthisissueinabit.)
TheWHEREclauseusescomparisonoperatorstocomparetwoexpressions
toeachother.SeeTable4-1forafulllistofthecomparisonoperators.The
operatorsarerelativelyself-explanatory,exceptforIN,LIKE,andIS,allof
whichwe’lldiscussinupcomingsections.
TABLE4-1 ComparisonOperators
ComparingDataTypes
Withinasinglecomparison,bothexpressionsshouldbethesamedatatypefor
thecomparisontowork.Thereareessentiallythreegeneralcategoriesofdata
typestoconsiderwhenperformingcomparisons—numeric,characterstring,and
date.Table4-2liststherulesforcomparingdatatypes.
TABLE4-2 RulesforDataTypeComparisons
YoumayhavenoticedthatIsaiddatatypes“should”bethesamefortwo
expressionsthatarecomparedtoeachother.Isay“should”becausethisisn’tan
absoluterule.Sometimesyoucangetawaywithcomparingexpressionsof
differentdatatypesprovidedthatSQLhasenoughinformationtoperforman
automaticdatatypeconversionandthereforetreatbothsidesofthecomparison
asthoughtheywerethesamedatatype,eventhoughtheyarenot.Whilethiscan
work,it’snotrecommended.Theresultsofsuchautomaticdatatypeconversions
canbeabittrickyandrelativelyunpredictable,soit’sbesttoavoidsuch
situations.WhenwediscussSQLfunctionslater,you’llseesomefunctionsthat
canbeusedtoperformexplicitdatatypeconversions,whichisthebetterchoice
forSQLprofessionals.Inthemeantime,don’tdependonOracleSQL’s
automaticdatatypeconversioncapabilitiesunlessyoulovetolivedangerously
anddon’tmindiftherestofuslaughatyou.
Here’sanexamplethatcomparesstringvalues:
ThissamplewillshowallcolumnsinalltherowsintheEMPLOYEEStable
wherethevaluefortheLAST_NAMEcolumnisthecharacterstring'Smith'.
Notethattextsearchesarecase-sensitivebydefault.Inotherwords,thisisa
differentquery:
ThesearchforemployeeswithavalueintheLAST_NAMEcolumnof
'SMITH'willnotfindthesamerowsthathaveaLAST_NAMEvalueof'Smith'.
Ifyouwanttodoasearchonbothpossibilities,seeChapter6,wherewe’ll
discusshowtohandlesuchsituationsusingSQLfunctions.
Whencomparingdates,Ialwaysliketorememberthisrhyme:“laterdatesare
greaterdates.”That’smytrickforrememberinghowtherulesofdate
comparisonwork.
LIKE
TheLIKEcomparisonoperatorisusefulforperformingwildcardsearcheson
characterdata.Ituseswildcardcharactersthatyoucanembedwithinatext
string.LIKEworkswithcolumnsofdatatypeCHARanddatatype
VARCHAR2.Technicallyitdoesn’treallyworkonDATE,butonapractical
levelitdoes—itperformsanautomaticdatatypeconversionoftheDATEvalues
involvedbeforeperformingthecomparison.
Thetwowildcardsymbolsare
Theunderscore(_),representingasinglecharacter
Thepercentsign(%),representingzeroormorecharacters
Youusetheunderscorewhenyouarereferencingafixedlengthofwildcard
characters.Underscorescanberepeatedasrequired.Forexample,thisqueryis
lookingforvaluesinthePORT_NAMEcolumnthatstartwiththestring'San',
followedbyablankspace,followedbyanyfourcharacters:
Incaseyoucan’ttell,that’sfourunderscoresafter'San'inthepreceding
query.Ifyouweretorunthisqueryagainstrowswiththesevalues:
thequerywillreturnonlythisvalue:
SanJuan
That’sbecauseofthefourunderscoresinthequery(line3),whichspecifically
askforfourunknowncharactersafter'San'.Nomore,noless.
Ifyouwanttoindicateanynumberofunknowncharacters,rangingfromzero
toinfinity,thenusethepercentsign.Thisquery,
willfindallthreerowsinthepreviousexample,likeso:
Thepercentsign,combinedwithLIKE,indicatesthatanynumberofcharacters
aresought.
Underscoresandpercentsignscanbeusedinanycombination,inanyorder,
inanylocationwithinthepattern.Here’sanexample:
ThisqueryislookingforvaluesinPORT_NAMEwithanyonerandom
characterinthefirstposition,followedbythelowercaseletterointhesecond
position,followedbyanywherefromzerotoanalmostinfinitenumberof
charactersafterthelettero.Thefollowingrowsmatchtherequest:
WhenworkingwithLIKE,youputthewildcardcharacter(orcharacters)
withinastringenclosedinsinglequotes.Thestringmustoccurafterthereserved
wordLIKE,notbefore.Inotherwords,thefollowingissyntacticallycorrectbut
doesn’tperformasyoumightthinkitshould:
Thisqueryisaskingwhetherthestringliteral'G_and%'happenstomatchthe
valuecontainedwithinPORT_NAME.Thisprobablyisn’twhatwasintended.
Thepointhereisthatthewildcardcharactersare“activated”onlyifthepattern
containingthemisontherightsideoftheLIKEreservedword.Thisisthequery
thatwasprobablyintended:
ThisquerywouldfindarowcontainingaPORT_NAMEvaluesuchasthis:
GrandCayman
Soremember,placethepatternafterLIKE,notbefore.Oraclewon’t
complainifyouscrewitup.Butyouroutputprobablywon’tbewhatyou’re
intending.
BooleanLogic
TheWHEREclauseincludessupportforBooleanlogic,wherebymultiple
expressionscanbeconnectedwithaseriesofBooleanoperators.Thissection
looksatthoseoperators,whattheyare,howtheyareused,andtheirorderof
precedence.
AND,OR
MostWHEREconditionsinvolvemorethanjustonecomparisonoftwo
expressions.MostWHEREclausescontainseveralsuchcomparisons.Thisis
wheretheBooleanoperatorscomein.Twoormorecomparisonsofexpressions
canbeconnectedbyusingvariouscombinationsoftheBooleanoperatorsAND
andOR.There’salsoathirdoperator—NOT—anditcanbeusedtoinvertan
ANDorORcondition.Booleanoperatorsevaluatemultiplecomparison
expressionsandproduceasingletrueorfalseconclusionfromtheseriesof
comparisons.Here’sanexample:
Let’sbreakthisdownthewayOracleSQLdoes.Considerthefollowingdata
listing:
Foreachrow,theWHEREconditionwilldothefollowing:
DeterminewhethertheSHIP_IDvalueequals3
DeterminewhethertheSTATUSvalueisequaltothestring'Pending'
Theresultsforeachrowareasfollows(Note:theitalicizedcolumnsarenot
outputlistings,butananalysisofBooleanoperatorcalculations):
Nowlet’sapplytheANDoperatortoeachrow:
TherulesofBooleanoperatorevaluationarethesameastheyarein
conventionalmathematics.SeeTable4-3foralistingofallthepossibleresults
ofBooleanoperatorexpressions.
TABLE4-3 BooleanExpressionCombinationsandResults
TherulesforBooleansare
ForAND,bothexpressionsmustbetrueforthecombinationtobe
true.Otherwise,theanswerisfalse.
ForOR,atleastoneexpressionneedstobetrueforthecombination
toevaluatetotrue.Otherwise,theanswerisfalse.
ThebasicsyntaxforaSELECTstatementwithaWHEREclausethat
includesBooleansisasfollows:
Lines4and6representthesamething—acomparisonoftwoexpressions.
AsingleWHEREclausemayincludeasmanyofthesecomparisonsasare
required,indicatedonline4andalsoonline6—providedtheyareeach
separatedbyaBooleanoperator.
NOT
ThereservedwordNOTispartofthesetofBooleanoperators.Itcanbeplaced
infrontofanexpressiontoreverseitsconclusionfromtruetofalse,orvice
versa.
Forexample,let’smodifyaSELECTstatementyousawearlier:
InthisSELECT,we’veaddedthereservedwordNOTonline5toreversethe
findingsofthecomparisonofthestring'Pending'tothevaluesinthecolumn
STATUS.IfyouweretorunthisversionoftheSELECTagainstthesamethree
rowsyouusedearlier,you’dgetaverydifferentresult:norowsreturned.Hereis
why:
WithanANDoperatorstillinusehere,nowourSELECTstatementwillreturn
norowssinceANDrequiresbothsidestobetrueandnorowssatisfythis
criteria.
Let’slookatanotherexample:
Asyoucanseefromthisexample,NOTcanbeusedwithoutanyotherBoolean
operators.Whilewe’reatit,theuseofNOTwejustconsideredhasthesame
effectasthis:
orasthis:
Both<>and!=inthesecontextshavethesameeffectastheuseoftheNOT
operator.
Ifyou’reanexperiencedprogrammerandhaveusedlanguagessuchas
Oracle’sPL/SQL,thissectionmayhaveyouwonderingwheretheBOOLEAN
datatypefitsintoSQL.Itdoesn’t.ThereisnoBOOLEANdatatypeinSQL.
ThereisinPL/SQL,butnotinSQL.Instead,expressionsarecomparedto
eachothertodetermineaBooleanconditionofTRUEorFALSE,andthe
Booleanoperatorscomparethemtodetermineananswer.Theconceptsof
TRUEandFALSEaresignificantthroughoutSQL,asweseewiththe
WHEREcondition.ButtherearenospecificdatatypesthatrepresentBoolean
values.
OperatorPrecedence
Justasthereisasetofrulesregardingtheorderofevaluatingarithmetic
operatorswithinexpressions,sotooarethererulesforevaluatingBoolean
operators.It’simportantthatyouremembertheorderinwhichSQLevaluates
Booleanoperators.
Thebottomline:NOTisevaluatedfirst.Afterthat,ANDisevaluatedbefore
OR.
Forexample,considerthefollowingdatalistingforatablecalled
SHIP_CABINS:
NowconsiderthisSQLstatementagainstthedatalisting:
Howmanyrowsdoyouthinkthisquerywillretrieve?Areyouthinking...
threerows,byanychance?Ifyouare,you’renotalone;mostpeopletendto
thinkthat.But,you’llreallygetfiverows.Why?Becauseoftherulesof
Booleanoperatorprecedence.AnEnglish-speakingpersonmightreadthatquery
asaskingforallrowswithaSTYLEvalueofeither'Suite'or'Stateroom'and
alsowithavalueforWINDOWof'Ocean'.ButSQLseesthisdifferently—it
firstevaluatestheANDexpression.SQLislookingforalltherowswhere
STYLEis'Stateroom'andWINDOWis'Ocean'...OR....anyrowwitha
STYLEvalueof'Suite',regardlessofitsvalueforWINDOW.Inotherwords,
it’sdoingthis:
Remember,onlyonesideinanORmustbeTRUE.Andgiventhiscriterion,
allfiverowswillevaluatetotrue.
YoucanuseparenthesestooverridetherulesofBooleanoperatorprecedence,
likethis:
Thatquerywillretrievethreerows.
AdditionalWHEREClauseFeatures
TheWHEREclauseofferssomeoptionstostreamlinethereadabilityofyour
code.ThissectiondescribesfeaturesthatareimportanttoadvancedWHERE
clauseusage.
IN
Sometimesyou’llfindyourselfcomparingasinglecolumntoaseriesofvarious
values.Here’sanexample:
Thatqueryiscorrect,butinsuchsituations,youhavetheoptionofchoosing
adifferentstyle.YoumaychoosetousethereservedwordINasanalternative.
Here’sanexample:
TherulesthatgoverntheuseoftheINoperatorincludethefollowing:
INcanbeusedwithdates,numbers,ortextexpressions.
Thelistofexpressionsmustbeenclosedinasetofparentheses.
Thelistofexpressionsmustbeofthesamedatatype—orbesimilar
enoughthatOraclecanperformautomaticdatatypeconversiontomake
themallthesame.
Thelistcanincludeanywherefromoneexpressiontoseveral,each
separatedbycommas.
Inaddition,theBooleanoperatorNOTmayprecedetheINoperator,as
follows:
TheuseofNOTwillidentifyrowsastrueiftheydonotcontainavaluefromthe
list.
You’llseelaterthatthereservedwordINisparticularlyimportantinthe
WHEREclausewhenwe’reusingsubqueries,whichwe’lldiscussinChapter9.
BETWEEN
Inadditiontotheformatswe’veseensofar,theWHEREclausesupportsa
featurethatcancompareasingleexpressiontoarangeofvalues.Thistechnique
involvesthereservedwordBETWEEN.Here’sanexample:
Thisistheequivalentofthefollowingstatement:
NoticethatBETWEENisinclusive.Inotherwords,itdoesn’tsimplylookfor
values“between”thetwocomparisonexpressions,butalsoincludesvaluesthat
areequaltothecomparisonexpressions.
Therangecanbespecifiedusinganyvalidexpression.
Therangeshouldbefromlowesttohighest.Ifyouspecifythehighervalue
firstandthelowervaluesecond,yourcodewillbeacceptedsyntacticallybut
willalwaysreturnzerorows.
TheNOTkeywordmayalsobecombinedwithBETWEEN.Thefollowing
arevalidstatements:
Thesetwoexamplesareequivalenttoeachother.
ISNULL,ISNOTNULL
RememberthattheNULLvaluerepresentsanunknownvalue.NULListhe
equivalentof“Idon’tknow.”Anyvaluethat’scomparedto“Idon’tknow”is
goingtoproduceanunknownresult—inotherwords,NULL.Here’san
example:
ThisSQLstatementwillneverretrieveanyrows,neverever,nevereverever.
Don’tbelieveme?Tryit,I’llwait.
Toldyou.Itwillnoteverretrieveanyrows,notevenifthevaluefor
CAPACITYisNULLwithinagivenrow.ThereasonisthatthisisaskingSQL
tocompareCAPACITYtoavalueof“Idon’tknow.”Butwhatifthevalueof
CAPACITYinthedatabasetableisactuallyNULL?Shouldn’tthisworkthen?
Shouldn’tyoubeabletoaskifCAPACITY=NULL?Whyisn’ttheexpression
NULL=NULLtrue?
Well,letmeaskyouthis:I’mthinkingoftwonumbers,andI’mnotgoingto
tellyouwhateitheroneofthemis.Instead,I’mjustgoingtoaskyouthis:are
thesetwonumbersequaltoeachother?Well?Arethey?
Ofcourseyoucan’tpossiblyknow.Ithinkyou’dhavetosaytheansweris“I
don’tknow,”orNULL.AndNULL,intermsofBooleanlogic,isalways
assumedtobeFALSE.So,anytimeyoucreateaWHEREconditionthatendsup
as“anything”=NULL,theanswerwillalwaysbeFALSE,andyou’llnevergeta
row—evenifthe“anything”isNULLitself.
Butwhatdoyoudowhenyoureallyneedtotestthevalueofsomethinglike
CAPACITYtodeterminewhetheritisNULLornot?Thereisananswer,andit’s
theSQLcomparisonconditionISNULLanditscompanionISNOTNULL.
Let’sredoourSELECTstatement:
Nowyou’reaskingforrowsfromthePORTStablewherethevaluefor
CAPACITYisunknowninthedatabase—whichiswhatISNULLasksfor.
That’saverydifferentquerythanaskingifCAPACITYhappenstobeidentical
tosomenumberthatwehaven’tidentifiedyet—whichiswhat=NULLasks.
TheoppositeoftheISNULLcomparisonoperatorisISNOTNULL,like
this:
Inthisquery,you’reaskingforallrowsinwhichtheCAPACITYvalueis
identifiedinthedatabase,inotherwords,whereveritisnotNULL.
Theseconceptsareimportant.Thisisyetanotherexampleofoneofthemany
waysinwhichSQLcodemightappeartobecorrect,mightexecutewithoutany
syntaxorexecutionerrors,andyetcanbetotallywrong.
Besureyougetthisright.Don’tscrewitup.Ifyoudo,theproblemswillbe
subtleandpotentiallydisastrousandmightbecomeapparentonlymonthslater
afterbaddataandincorrectreportshavebeencirculated.So,remember,never
usethis:
=NULL
Thereisneveragoodreasontousethat,ever.Alwaysusethisinstead:
ISNULL
Gotit?
AdditionalConcepts
TherearemorewaystocustomizeaWHEREclausethanwhathasbeen
presentedhere.Upcomingtopicsthatwillbeaddressedinthisbook—andtested
ontheexam—includethefollowing:
Subqueries
Setoperators
TheseandotherissuesareimportanttotheWHEREclauseandimportantto
theexam.Theywillbecoveredontheirowninupcomingchapters.
CERTIFICATIONOBJECTIVE4.03
UseAmpersandSubstitutiontoRestrictandSort
OutputatRunTime
Theinclusionoftheampersandsubstitutionvariableasanexamobjectiveisa
bitunusualinthatitrepresentsaSQL*Plusfeaturetobeaddressed.Allother
examobjectivesfocusonOracle’simplementationoftheSQLlanguage,
includingANSIstandardSQLaswellasOracle’senhancements,suchasthe
DECODEfunction.Buttheampersandsubstitutionvariableisafeatureof
SQL*Plusexclusively.SQL*PlusisatoolforworkingwithOracle’srelational
databasebyusingSQLcommandsinteractivelyorviabatchprocessing,andit
offersseveraluniquecommands,systemvariables,andothercapabilities,allof
whichareuniquetoSQL*Plus.Theampersandsubstitutionvariableisoneof
thoseSQL*Plusfeatures,butitisnotpartoftheSQLlanguage.Youwon’tfinda
descriptionofitintheOracleSQLLanguageReferencemanual,not,atleast,at
thetimeofthiswriting.ButyouwillfinditdescribedinOracle’sSQL*Plus
documentation.
ThepurposeofampersandsubstitutionistoprovideaSQLscriptwitharun-
timeparameterfeaturewhenusedwithinSQL*Plussessions.Thevalueofthe
parametercanbesetviabatchbeforeorinteractivelyduringagivenSQLscript’s
execution.Eitherway,thepointistodecouplethevalueoftheparameterfrom
thescript,empoweringanindividualscripttoexecuteinavarietyofways.
Furthermore,thesubstitutionvariablecanrepresentanyaspectoftheSQL
script.ItcanbeusedtospecifyavalueinaWHEREclause,acolumnina
SELECTstatement,thenameofatable,orvirtuallyanyportionofthecode
withinascript,includingkeywords,SQLcommands,ordataofanykind,such
asinputdataforanINSERTorUPDATEstatement,thecriteriaofaHAVINGor
GROUPBY,oranything.
Tofullyleveragethepoweroftheampersandsubstitutionvariable,youneed
toknowmorethanthevariable.Thereareavarietyofsurroundingcommands
youshouldunderstandtoconfigureandimplementtheampersandsubstitution
variable.Thissectiondescribesthosesupportingfeatures,allofwhichmay
appearontheexam.
Inthissection,we’lllookattheampersandsubstitutionvariableitselfand
howtouseitduringexecution.Nextwe’llpredefineavalueusingtheSQL*Plus
commandsDEFINEandUNDEFINE.Thenwe’lldiscusshowSQL*Plussystem
variableswork,afterwhichwewilllookattwoofimportanceforourdiscussion:
theVERIFYsystemvariableandtheDEFINEsystemvariable(nottobe
confusedwiththeDEFINEcommand).Finally,we’lldiscusstwoadditional
SQL*Pluscommandsthatyoucanusewithsubstitutionvariables:ACCEPTand
PROMPT.
&
Let’slookatanexampleoftheampersandsubstitutionvariablebyusingthe
SHIP_CABINStable.Let’ssaythetablecontainsthefollowingrowsofdata:
HereisascriptthatwillquerySHIP_CABINSanduseanampersand
substitutionvariableontheendofthethirdlinetoenableustosettheWHERE
criteriaatexecutiontime:
Oncewehavecreatedthisscript,wecanexecuteitandspecifythevaluefor
ROOM_NUMBER.RefertoFigure4-7toseewhathappenswhenweexecute
thisscriptwithinaSQL*Plusclient.
FIGURE4-7 Usinganampersandvariabletospecifyroomnumberatrun
time
Duringexecution,thepresenceoftheampersandsubstitutionvariabletriggers
an“Entervalueforrno:”requestfromthesystem.Inresponse,wetype104and
pressENTER.Thesystemthendisplaystheoldline3,aswellasthenewline3
showinghowthevaluefor“rno”hasbeenreplacedwith104.Finally,the
statementisexecutedasthough104hadbeenintheplaceof“rno”allalong,and
weseetheresult:areturnofrowswhereROOM_NUMBER=104.
Theampersandsubstitutionvariableconsistsofthefollowing:
Thesubstitutionvariableprefix(anampersand)
Thesubstitutionvariablename,consistingofanysubsequent
contiguoussetofalphanumericcharactersuntilthenextblankspace
Thevariablenameyouspecifywillberepeatedatruntimeinthecontextofa
question,asyoucanseeintheexample.IusedthenameRNO,butyoucould
haveusedsomethingmoredescriptive,perhapsRoom_Numberorsomething.
Inourexample,weusedthesubstitutionvariabletoprovideavaluefora
numericcolumnatruntime.Butwhatabouttextvalues?Thesubstitution
variablecanbespecifiedwithinquotesintheSQLstatement.Here’sanexample:
Let’sseewhathappenswhenweexecutethiswithinaSQL*Plusclient;see
Figure4-8.
FIGURE4-8 Usinganampersandvariablewithquotes
Asyoucanseeintheexample,thesubstitutionvariableusedfromwithin
quotesworksjustthesame.Alternatively,youcanomitthequotesintheSQL
andchooseinsteadtoprovidethequotesatruntime(seeFigure4-9).
FIGURE4-9 Usinganampersandvariablewithoutquotes
YoucanreplacevirtuallyanysectionofSQLcodewithasubstitution
variable;seeFigure4-10.
FIGURE4-10 UsinganampersandvariablewithSQLcolumnspecifications
Notehowwejustusedasinglesubstitutionvariableintheplaceofacolumn
listoftheSELECTstatement.Atruntime,whenprompted,wetypedinthe
namesoftwocolumns,separatedwithacomma.Thesetwonamesare
substitutedatruntime,andthecodeexecutesaccordingly.
Sofarwehaveworkedwiththesubstitutionvariablebyspecifyingitsvalue
atruntime.Butyoucanpredefineitaswell,usingtheSQL*Pluscommand
DEFINE.
DEFINEandUNDEFINECommands
Youcanuseasinglestatementonitsowntodefineasubstitutionvariableusing
theDEFINEstatementsothatwhenSQLcodewithasubstitutionvariableis
executed,thevalueyou’vealreadydefinedwillbeused.Here’sanexample:
DEFINEvWindows=Ocean
Notethatweomitthesubstitutionvariableprefix;thereisnoampersandat
thebeginningofthevWindowssubstitutionvariable.
Tolistallexistingdefinedvariablesforagivensession,useDEFINEona
linebyitself,likethis:
DEFINE
Alsonote:WechosetoexecutetheSQL*PluscommandDEFINEwithouta
semicolonattheend.SQLcommandsrequireasemicolonattheend;SQL*Plus
commandsdonotrequirethem,yettheydoacceptthem.Soyoumaychooseto
includeoromitthem.SQLcommandsrequiresemicolons.ButforSQL*Plus
commands,semicolonsareoptional.
Thescopeofthesubstitutionvariableisthesession.Oncedefined,the
variablewillpersistthroughthesessionoruntilitis“undefined.”
UNDEFINEvWindows
WithoutanUNDEFINE,thevWindowssubstitutionvariablewillpersistfor
thedurationoftheloginsession.WewilluseDEFINEwithaseriesofSQL
statementstoseehowitworks.Butfirstlet’salsolookatSETandSHOW.
TheSETandSHOWCommands
TheampersandsubstitutionvariableisaSQL*Pluscommand,anditsbehavioris
controlledbythesettingsofsomeoftheSQL*Plussystemvariables.SQL*Plus
includesanumberofsystemvariablesthatcanbeconfiguredusingtheSET
command.Thecurrentconfigurationofasystemvariablecanbedisplayedusing
theSHOWcommand.Foranexample,seeFigure4-11.
FIGURE4-11 SHOW
SHOWcanbeusedwithanyoftheSQL*Plussystemvariablestodisplaythe
particularsystemvariable’scurrentstate.Agivensystemvariable’scurrentstate
consistsofwhetheritisONorOFFandoptionallymayincludeconfiguration
detailsuniquetothesystemvariable.
Toseethefullsetofallsystemvariablesandtheirstate,useSHOWALL.
Forthepurposesofworkingwiththesubstitutionvariable,weareprimarily
interestedintwoparticularsubstitutionvariables:VERIFYandDEFINE.Let’s
lookatbothinmoredetail.
TheVERIFYSystemVariable
NotehowSQLhasbeeninformingusofthe“old”and“new”linesofcodefor
eachsubstitutionvariable.ThiscanbeturnedonoroffusingaSETVERIFY
statement.
SETVERIFYOFF
Onceoff,the“old”and“new”lineswillnolongerdisplay.Alternatively,you
canuseanabbreviatedversionofVERIFY.
SETVEROFF
TorestoreVERIFY,turnitbackon.
SETVERIFYON
SettingVERIFYbacktoONwillensurethe“old”and“new”linesappear
whenusingadefinedvariableinyourSQLscript.
ACCEPTandPROMPT
TwoadditionalSQL*Pluscommandsareusefulwhenworkingwithsubstitution
variables.PROMPTisacommandtodisplayaninteractivemessagetoanend
user.ACCEPTwillreceivedatafromausertostoreinapredefinedvariable.
Together,thesetwocommandsarecommonlyusedininteractivescriptswhere
SQL*Plussubstitutionvariablesareinuse.
Foranexample,let’spromptanendusertotypeinanumber,acceptthe
typednumberasthevalueforavariablewe’llnamevRoomNumber,andthen
incorporatethevariableintoaSQLscript.
Next,we’llstoretheSQLscriptinasimpletextfilethatwe’llname
The_Script
.sql.We’llstorethefileinthedefaultfolderofourlocalSQL*Plusclientand
executethefile.SeeFigure4-12fortheresults.
FIGURE4-12 ACCEPTandPROMPT
NotethatPROMPTdoesnotrequirequoteswheninvokedasastand-alone
statementbutdoesrequirethemwhenusedinconjunctionwithACCEPT.
Inourexample,theenduserprovidesaroomnumberof104,andthe
subsequentSQLscriptreturnstheappropriaterow.UsingPROMPTyoucan
createamoremeaningfulquestionoftheenduser.BycombiningPROMPTwith
ACCEPTyoucanacceptananswerfromtheenduserandstorethevalueina
variableyoureferencelaterintheSQLscript.
Animportantnote:Weincludedaperiodattheendofourdefinedvariable,
yettheperioddidnotdisplayinteractively.Hadweaddedaspaceafterthe
definedvariableandbeforetheperiod,theperiodwouldhavedisplayed.Forthat
matter,anyadditionaltextafterthedefinedvariablewilldisplay,provideditis
separatedfromthedefinedvariablewithoneormorespaces.
Youcanuseanysimpletexteditortocreatebatchscriptsforprocessingby
SQL*Plus.IuseNotepadorUltraEdit.Youshouldsavethescriptinafile
endingwiththe*.SQLsuffix,butthisisnotrequired.Forexample,ifyou
createafilecalledRun.sqlandthensaveitonyouroperatingsystem
somewhere,youcaninvokeitfromwithinSQL*Plusbytyping@[folder
location]Run,where@tellsSQL*Plustoexecutethescript,[folderlocation]is
thelocationofthefile,RUNisthefileprefix,andthe.SQLfilesuffixis
assumed.Alternatively,youcanuse@[folderlocation]Run.SQL.Thespecific
folderlocationisimplementationspecific,andyoushouldcheckwithyour
localdatabaseadministratorand/ordocumentationtodetermineitslocation.
TheDEFINESystemVariable
Tousethesubstitutionvariable,theDEFINEsystemvariablemustbesettoON.
ThedefaultsettingforDEFINEisONforagivensession,soyougenerallydon’t
needtospecificallysetit.Buttobesure,youcanensurethatDEFINEisON
beforeyouusethesubstitutionvariablebyexecutingthefollowingSQL*Plus
statementpriortoanyuseofasubstitutionvariable:
SETDEFINEON
Thisisparticularlyimportantifyouarecreatingabatchscriptandyouwon’t
necessarilyknowatthetimethebatchscriptisexecutedwhetherthedefaultON
valueofDEFINEisset.ItmightbeturnedOFF.SettingittoONexplicitlywhen
itisalreadyONwillnottriggeranycomplaintsfromthesystem.So,itis
commonpracticetouseSETDEFINEONatthebeginningofabatchscriptthat
usesit.
TosetDEFINEtoOFF,usethis:
SETDEFINEOFF
NotethatSQL*Plusstatementsdonotrequireasemicolonattheend.You
canusethemifyouwant,buttheyarenotrequired.
Theexamobjectivesspecificallyrefertothesubstitutionvariableasthe
“ampersandsubstitutionvariable,”buttheampersanditselfcanbechangedwith
avariationoftheSETDEFINEcommandthatdoesn’tuseeitherONorOFFbut
insteadresetsthesubstitutionvariableprefixcharacter.Here’sanexample:
SETDEFINE*
ThisSETDEFINE*statementwillchangethesubstitutionvariableprefixto
anasterisk.Let’sseeitinactionwithacombinationofstatements.Forthis
example,we’llsetasystemvariablewehaven’tusedyet,ECHO.SettingECHO
toONisusefulwhenexecutingaseriesofstatementsinbatch;ECHOensures
ourstatementsthemselveswilldisplay,notjusttheiroutput,sowecanseethe
fulleffectofourscript.
Next,we’llsetVERIFYtoOFFsothe“old”and“new”lineswon’tdisplay
wheneverweinvokeasubstitutionvariable.
Next,we’llsetDEFINEtoOFFanduseSHOWtoconfirmthatDEFINEhas
infactbeenturnedoff.Butthenwe’lluseSETDEFINEtochangetheprefix
fromthedefault(oftenanampersand)totheasterisk.ThenwewilluseSHOW
DEFINEagaintoseethatthedefaultprefixhasinfactbeenresetandthatthe
DEFINEsystemvariableisnolongerOFF;changingthedefaultprefixhasthe
by-productofalsoturningtheDEFINEsystemvariablebacktoON.
Next,we’lltrytouseasubstitutionvariablewiththeampersand,butsince
we’vechangedtheprefixtosomethingelse,theampersandwillnolongerbe
recognizedandinsteadwillbetreatedasstandardtext.
Finally,we’llrepeatthesameSQLscript,butusingthenewlyspecified
substitutionvariable’sprefixofanasterisk,andwe’llseethattheSQLstatement
producestheoutputwe’reexpecting.
FortheresultsofallofthiswhenexecutedasabatchscriptwithinSQL*Plus,
seeFigure4-13.
FIGURE4-13 Changingtheprefixofthesubstitutionvariable
Notethatthesubstitutionvariableprefixislimitedtoonecharacteronly.
Also,ifDEFINEissettoOFFandyouchangetheprefix,youwillalso
automaticallysetDEFINEtoON.
CERTIFICATIONOBJECTIVE4.04
UsetheSQLRowLimitingClause
YoucanusetheSQLrowlimitingclausetoreturnonlyasubsetofthatdata.The
SQLrowlimitingclauseisnewwithOracle12c.Itcanbeusedafterany
WHEREorORDERBYclausetolimitthenumberorpercentageofrowstobe
returnedbyaquery.Granted,allWHEREcriteriaarealreadyintendedtolimit
returneddatafromalogicalstandpoint—that’stheentirepurposeoftheWHERE
clause.Buttherowlimitingclausetakesamuchbroaderview.Whereasother
aspectsofyourqueryaregenerallyfocusedonbusinesslogic,therowlimiting
clausecanbeusedtochoose,forexample,onlythefirst10rowsofquerieddata,
theseconddozenrows,orthelast50rows.
FETCH
LimitsontherowsreturnedbyaSELECTstatementarespecifiedbytheFETCH
statement.HereisanexampleofasingleSELECTstatementusingtheSQLrow
limitingclause:
ThekeywordFETCHsignifiesthepresenceoftherowlimitingclause.Inthis
example,thequerywillreturnuptoeightrowsandnomore.Forexample,if
onlysixrowsexist,thenallsixwillbereturned.If500rowsexist,only8willbe
returned.
FETCHcanspecifyanumberofrows,aswejustsaw,orapercentage,like
so:
TheFETCHclauseconsistsofthefollowingrequiredkeywordsandoptions:
ThekeywordFETCH(required).
EitherthekeywordFIRSTorthekeywordNEXT(oneisrequired).
Eitherisacceptable,andthereisnofunctionaldifferencebetweenthe
two;theoptionisavailableforgrammaticalreadabilityonly.
Avalidexpressionthatevaluatestoanumber(optional).
IfnonumericvalueispresentbuttherestoftheFETCH
keywordsareinplace,FETCHwilldefaulttoonerow.
AnoptionalkeywordPERCENTiftheimmediatelypreceding
numberistobeinterpretedasapercentageinsteadofanumeric
value.PERCENTwillneverreturnapartialrow,andpercentagesare
roundeduptothenearestnumberofrows.Forexample,ifyou
specifyafetchof1PERCENToffouravailablerows,onerowwill
bereturned.Ifyouspecify26PERCENToffouravailablerows,two
rowswillbereturned.
EitherthekeywordROWorthekeywordROWS(oneisrequired).
JustliketheFIRST/NEXTkeywords,eitherROWorROWSis
acceptable,andthereisnofunctionaldifferencebetweenthetwo;the
optionisavailableforgrammaticalreadabilityonly.
EitherthekeywordONLYorthekeywordsWITHTIES(oneis
required).
ONLYwillreturnthenumberorpercentageofrowsspecified
andnomore.
WITHTIESisbestunderstoodwithanexample,butinshort,
WITHTIESwillreturneverythingreturnedbyONLYbutpossibly
withsomeadditionalrows.Thoseadditionalrowswillbeconsidered
onlyifanORDERBYclauseisincludedintheSELECTandthat
ORDERBYsortsrowsinsuchawaythattherearerowsremaining
thatwerelogicallyconsistentwiththeparticularORDERBYclause
butarbitrarilyomittedtosatisfytherowlimit.Thisrequiressome
explanationandisbestillustratedwithanexample;seethenext
section.
WITHTIES
WITHTIESis“tied”totheORDERBYclause.WithoutORDERBY,WITH
TIESwillbehavethesameasONLY.
Let’sconsideranexampleofWITHTIES.Let’ssayyouhaveeightrowsina
tableORDERS,fiveofwhichhaveavalueintheLINE_ITEMScolumnof2.If
yousorttherowsbyLINE_ITEMSandthenFETCHFIRST50PERCENT
ROWSONLY,you’llgetthefirstfourrowsreturnedbytheORDERBYclause,
asshowninFigure4-14.
FIGURE4-14 SELECTwithFETCH
Thatmakessensebecausethereareeightrowsintotal,so50percent—half—
isfourrows.However,yourORDERBYclausesortedrowsbyLINE_ITEMS,
andthenumberofrowswithaLINE_ITEMvalueof2doesn’tconveniently
breakat50percent,sothereareadditionalrowswithaLINE_ITEMof2that
didn’tmakethecut.
There’snothinginherentlyrightorwrongaboutthatapproach.Ifyour
objectiveistoonlyget50percentoftherowsanditdoesn’tmattertoyour
businessobjectivesthatthelinewasarbitrarilydrawnthroughthesetofrows
whereLINE_ITEM=2,thenyou’refine.
Butwhatifyourobjectiverequiresyoutoincludethoseadditionalrows
whereLINE_ITEM=2?Ifyourobjectiveislessconcernedwithanarbitrary
breakpointandmoreconcernedwithincludingalogicallyconsistentsetofrows,
youcanusetheWITHTIESclause.
So,let’sedittherowlimitingclausetoinsteadbeFETCH50PERCENT
ROWSWITHTIES.Theresult:we’llgetthesame50percentreturnedinthe
immediatelyprecedingexampleplusanyadditionalrowswiththesamevalue
specifiedintheORDERBYcriteria,asshowninFigure4-15.
FIGURE4-15 SELECTwithFETCH…WITHTIES
NotethatwhenwecombineORDERBYandaFETCH…WITHTIES,we
endupwiththefirst50percentplusallrowsthatlogicallyequatetothe
LINE_ITEMvalueatthecutoffpoint.
WesaidthatWITHTIESis“tied”totheORDERBYclause.Whatifwe
omittedtheORDERBYclause?SeeFigure4-16.
FIGURE4-16 SELECTwithFETCH…WITHTIES,noORDERBY
WithoutORDERBY,WITHTIEShasnoeffect.WithoutORDERBY,the
useofWITHTIESproducesthesameresultasifwehadexecutedFETCH…
ONLY.Wedidn’tgetasyntaxerror.Wesimplyreceivedthesameresultsasif
we’dusedFETCH…ONLY.
OFFSET
Bydefault,therowlimitingclausewillstartwiththefirstreturnedrow,butyou
canoptionallyspecifythattheFETCHstartelsewhere,atsomespecificnumber
ofrowsintotherange.Here’sanexample:
Inthisexample,theFETCHwillskipthefirstfiverows,startatthesixthrow,
andreturnthesixthandseventhrows,foratotaloftworowsreturned.
Thefollowingtablelistswhathappenswithvariousspecificationsof
OFFSET:
CERTIFICATIONSUMMARY
TheORDERBYclauseoftheSELECTstatementisthemethodforsortingthe
rowsofoutputreturnedbytheSELECTstatement.TheORDERBYclauseis
optional,butifincludedinaSELECTstatement,itisalwaysthefinalclauseof
theSELECTstatement.
ORDERBYspecifiesoneormoreexpressions.Eachexpressionisusedby
ORDERBYtosorteachrowintheresultset.Outputrowsaresortedfirstbythe
firstexpression,andforanyrowsthatsharethesamevalueforthefirst
expression,thoserowswillbesubsortedforthesecondexpression,andsoon.
ExpressionsforORDERBYfollowthesamerulesasexpressionsinthe
SELECTstatement’sselectlistandtheWHEREclause.Eachexpressionshould
ideallyreferenceacolumninthetable,butthisisn’trequired.
TheASCandDESCreservedwordscanbeusedinanORDERBYclauseto
determinethedirectionofsortingforeachindividualexpression.ASCwillsort
valuesinascendingorder,andDESCwillsortindescendingorder.ASCisthe
defaultanddoesnotneedtobespecified.ThespecificationofASCorDESC
shouldfolloweachexpressionwithaspaceseparatingthem.
ORDERBYsortsvaluesaccordingtodatatypes.Withnumericdatatypes,
lownumbersarelow,andhighnumbersarehigh;withdates,yesterdayislower
thantomorrow,andnextweekishigherthanlastmonth;withcharacters,Zis
lessthana,andthecharacterstring'10'islessthanthecharacterstring'3'.
ORDERBYcanspecifyexpressionsintheSELECTstatement’sselectlistby
referencingthecolumnalias,ifonewascreatedwithintheSELECTlist.
ORDERBYcanalsoidentifyexpressionsintheSELECTlistbythenumber
correspondingtothepositionoftheitemintheSELECTlist;forinstance,
ORDERBY1willsortbythefirstitemintheSELECTlist.
ORDERBYcancombineallofthesefeaturesintooneseriesoforderitems.
Acolumnalias,ifspecifiedintheselectlist,isnotrecognizedinthe
WHERE,GROUPBY,orHAVINGclause.
TheWHEREclauseispartoftheSQLstatementsSELECT,UPDATE,and
DELETE.ItisusedtoidentifytherowsthatwillbeaffectedbytheSQL
statement.FortheSELECTstatement,WHEREdetermineswhichrowsare
retrieved.FortheUPDATEstatement,itdetermineswhichrowswillbeupdated.
FortheDELETEstatement,itdetermineswhichrowswillbedeleted.The
WHEREclauseconcernsitselfwithentirerows,notjustthecolumnsthatarethe
subjectoftheparticularSQLstatementofwhichitisapart.TheWHEREclause
canreferencecolumnsthatarenotintheSELECTlist.
TheWHEREclauseisoptional.IfincludedinaSELECTstatement,itmust
followtheFROMclause.
TheWHEREclausecomparestwoexpressionsanddetermineswhetherthe
resultofthecomparisonistrueorfalse.Atleastoneofthetwoexpressions
shouldincludeacolumnfromwhatevertabletheSQLstatementisintendedto
addresssothattheWHEREclauseisrelevanttotheSQLstatement.The
expressionsarecomparedusingcomparisonoperators.Examplesincludethe
equalsign,thenot-equalsign,andthegreater-thanandless-thansigns.
AseriesofcomparisonscanbeconnectedusingtheBooleanoperatorsAND
andOR.TheNOToperatorcanalsobeincluded.Together,theseexpressionscan
formcomplexWHEREclauses.
TheLIKEoperatorcanactivatethewildcardcharacters.Thetwowildcard
charactersaretheunderscore(_)andthepercentsign(%).
TheINoperatorcancompareasinglevaluetoasetofvalues.Anexpression
usingtheINoperatorwillevaluatetotrueifthatsinglevaluematchesanyofthe
valuesintheexpressionlist.
WhenusingtheWHEREclausetolocaterowsthathaveaNULLvalue,never
usethe=NULLcomparison;instead,alwaysusetheISNULLorISNOTNULL
comparisonoperator.
TheampersandsubstitutionvariableisusedwithinSQL*Plustosupport
interactiveandbatchprocessing.Theampersandisaprefixthatcanbe
configureddifferentlywiththeSETDEFINEstatement.Thesubstitution
variablecanbeturnedonoroffwithSETDEFINEONorSETDEFINEOFF.
YoucanalsouseSETVERIFYtoturnonorofftheautomaticdisplayof
“old”and“new”linesofcodecontainingthespecifiedsubstitutionvariable
value.AnySQLscriptcanleveragethesubstitutionvariabletoreplaceanycode
orvaluewithinanysubsequentscript.
TherowlimitingclauseisnewtoOracle12c.Itspurposeistolimitthe
numberofrowsreturnedbyaSELECTstatement.Therowlimitingclauseis
specifiedwithFETCHandsyntacticallyfollowsORDERBY(ifitisincluded).
TheFETCHclausespecifiesarangeofrowstobereturnedasasubsetofthe
SELECTstatements.Therangecanbedeterminedbynumbersofrowsorasa
percentage.Ifnorangeisspecified,FETCHdefaultstoonerow.TheFETCH
keywordsFIRSTandNEXTbehavethesame,sodoROWandROWS;both
pairsareacceptableandexistforreadability.WITHTIEScanbeusedtoinclude
rowsthathaveidenticalsortedvaluesasspecifiedbyORDERBYandwould
otherwisebearbitrarilyomittedfromthereturnset.AnoptionalOFFSETvalue
canbeincludedtodeterminethestartingpointoftherange.OFFSETisspecified
bynumbersofrows(notapercentage).Ifomitted,OFFSETdefaultstozero,and
ifenteredasanegativenumber,OFFSETwillbetreatedasthoughzero.
✓TWO-MINUTEDRILL
SorttheRowsThatAreRetrievedbyaQuery
ORDERBYisanoptionalclauseusedtosorttherowsretrievedin
aSELECTstatement.
Ifused,ORDERBYisalwaysthelastclauseintheSELECT
statement.
ORDERBYusesexpressionstodirectthesortingorderoftheresult
setoftheSELECTstatement.
Eachexpressionisevaluatedinordersothatthefirstitemin
ORDERBYwilldotheinitialsortofoutputrows,theseconditemlisted
willsortanyrowsthatshareidenticaldataforthefirstORDERBY
element,andsoon.
ORDERBYcansortbycolumnsinthetable,regardlessofwhether
thecolumnsappearintheSELECTstatement’sselectlist.
ORDERBYcanalsosortbyexpressionsofanykind,followingthe
samerulesofexpressionsthatyou’veseenwiththeWHEREclauseand
theselectlist.
Numericdataissortedbydefaultinascendingorder,fromlower
numberstohigher.
Characterdataissortedbydefaultinascendingorder,fromAtoZ.
Datedataissortedbydefaultinascendingorder,frompriordatesto
laterdates.
Allsortsdefaulttoascendingorder,whichcanbespecifiedwiththe
optionalkeywordASC.
Sortordercanbechangedtodescendingorderwiththekeyword
DESC.
ORDERBYcanidentifycolumnsbycolumnaliasorbyposition
withintheSELECTlist.
LimittheRowsThatAreRetrievedbyaQuery
TheWHEREclausecomesaftertheFROMclause.
WHEREidentifieswhichrowsaretobeincludedintheSQL
statement.
WHEREisusedbySELECT,UPDATE,andDELETE.
WHEREisanoptionalclause.
ExpressionsformthebuildingblocksoftheWHEREclause.
Anexpressionmayincludereferencestocolumnnamesaswellas
literalvalues.TheWHEREclausecomparesexpressionstoeachother
usingcomparisonoperatorsanddetermineswhetherthecomparisonsare
trueorfalse.
Booleanoperatorsmayseparateeachcomparisontocreatea
complexseriesofevaluations.Collectively,thefinalresultforeachrow
inthetablewilleitherbetrueorfalse.Iftrue,therowisreturned;if
false,itisignored.
TheBooleanoperatorsareAND,OR,andNOT.
TherulesofBooleanoperatorprecedencerequirethatNOTbe
evaluatedfirst,thenAND,andthenOR.
ParenthesescanoverrideanyBooleanoperatorprecedence.
Whencomparingdatedatatypes,earlierdatevaluesareconsidered
“less”thanlaterdates,soanythinginJanuarywillbe“lessthan”
anythinginDecemberofthesameyear.
Whencomparingcharacterdatatypes,theletteraislessthanthe
letterz,uppercaselettersare“lowerthan”lowercaseletters,andthe
characterrepresentationof3isgreaterthanthecharacterrepresentation
of22,eventhoughtheresultswouldbedifferentiftheywerenumeric
datatypes.
LIKEcanbeusedtoactivatewildcardsearches.
INcanbeusedtocompareasingleexpressiontoasetofoneor
moreexpressions.
BETWEENcanbeusedtoseewhetheraparticularexpression’s
valueiswithinarangeofvalues.BETWEENisinclusive,notexclusive,
sothatBETWEEN2and3includesthenumbers2and3aspartofthe
range.
UseISNULLorISNOTNULLwhentestingacolumntosee
whetheritsvalueisNULL.
UseAmpersandSubstitutiontoRestrictandSortOutputatRun
Time
ThesubstitutionvariableisusedinbatchorinteractiveSQL*Plus
processing.
ItisnotaSQLfeature,butaSQL*Plusfeature.
ThesubstitutionvariablecanspecifyaWHEREclauseparameter,
ORDERBYoption,columnlist,tablename,anyportionoftheSQL
keyword,oranyportionoftextwithinaSQLscript.
ThesubstitutionvariableworkswithSELECT,INSERT,UPDATE,
orotherSQLstatements.
Theampersandcanberedefinedtoanalternativesinglecharacter.
VERIFYisasystemvariable;whensettoON,itensuresthatthe
useofasubstitutionvariablewilltriggerthedisplayofan“old”and
“new”lineoftext.
TheDEFINEsystemvariablespecifiestheprefixforusewitha
substitutionvariable.
ACCEPTcanbeusedtoacceptinteractivelyspecifiedvaluesfora
substitutionvariable.
PROMPTcanbeusedtoissuerichpromptquestionstoauser
duringaninteractiveSQL*Plussession.
UsetheSQLRowLimitingClause
FETCHisanoptionalclausethatfollowsWHEREandORDER
BY.
FETCHcanspecifyarangeintermsofanumberofrowsora
percentageofrows.
ExpressionscanbeusedinFETCHtospecifythenumberofrows
orpercentageofrows.
Ifnorangeisspecified,FETCHdefaultstoonerow.
AnexampleisFETCHFIRST20ROWSONLY;.
ThekeywordFIRSTcanbeexchangedwiththekeywordNEXT;
botharefunctionallyidentical,butoneofthetwokeywordsisrequired.
ThekeywordROWcanbeexchangedwiththekeywordROWS;
botharefunctionallyidentical,butoneofthetwokeywordsisrequired.
TheSQLrowlimitingclauserequireseitherthekeywordONLYor
thekeywordsWITHTIES.ONLYreturnsrowsthatmatchtheFETCH
criteriaonly.WITHTIESreturnsrowsthatmatchtheFETCHcriteria
andalsosatisfytheORDERBYclauseatthesameFETCHlogical
dividingpointintherange.
WithoutORDERBY,WITHTIESbehaveslikeONLY.
SELFTEST
Thefollowingquestionswillhelpyoumeasureyourunderstandingofthe
materialpresentedinthischapter.Chooseonecorrectanswerforeachquestion
unlessotherwisedirected.
SorttheRowsThatAreRetrievedbyaQuery
1.ReviewthisSELECTstatement:
Assumethatalltableandcolumnreferencesexistwithinthedatabase.
WhatcanbesaidofthisSELECTstatement?
A.TherowswillsortinorderbySHIP_IDandthenbyCAPACITY.
Allrowswillsortindescendingorder.
B.TherowswillsortinorderbySHIP_IDinascendingorderand
thenbyCAPACITYindescendingorder.
C.ThestatementwillfailtoexecutebecausetheORDERBYlist
includesacolumnthatisnotintheselectlist.
D.ThestatementwillfailtoexecutebecausethereisnoWHERE
clause.
2.ReviewthisSELECTstatement:
Assumealltableandcolumnreferencesexistinthedatabase.Whatcan
besaidofthisSELECTstatement?
A.Thestatementwillexecutesuccessfullyandasintended.
B.ThestatementwillexecutebutnotsortbecausetheORDERBY
clauseiswrong.
C.ThestatementwillfailtoexecutebecausetheORDERBY
clauseincludesthewordLIKE.
D.Noneoftheabove.
3.WhichifthefollowingistrueoftheORDERBYclause?(Choose
two.)
A.Itisoptional.
B.ItcanbeusedintheUPDATEstatementaswellasSELECTand
DELETE.
C.Itcansortrowsbasedondatathatisn’tdisplayedaspartofthe
SELECTstatement.
D.IfthelistofORDERBYexpressionsusesthe“byposition”
form,thenallexpressionsintheORDERBYmustusethe“by
position”form.
4.IfyouareusinganORDERBYtosortvaluesindescendingorder,in
whichorderwilltheyappear?
A.Ifthedatatypeisnumeric,thevalue400willappearfirstbefore
thevalue800.
B.Ifthedatatypeischaracter,thevalue'Michael'willappearfirst
beforethevalue'Jackson'.
C.Ifthedatatypeisdate,thevalueforJune25,2010,willappear
beforethevalueforAugust29,2010.
D.Ifthedatatypeischaracter,thevalue'130'willappearfirst
before'75'.
LimittheRowsThatAreRetrievedbyaQuery
5.ReviewthefollowingdatalistingforatableVENDORS:
NowreviewthefollowingSQLstatement:
HowmanyrowswilltheSELECTstatementreturn?
A.2
B.1
C.0
D.Nonebecauseitwillfailduetoasyntaxerror
6.ReviewthefollowingdatalistingforatablecalledSHIP_CABINS:
TheblankvaluesareNULL.NowreviewthefollowingSQLstatement
(linenumbersareaddedforreadability):
HowmanyrowswilltheSQLstatementretrieve?
A.0
B.1
C.2
D.Nonebecauseyoucannotuseparenthesesinline3tosurround
theexpressions
7.ReviewthefollowingdatalistingforatableSHIPS:
IntheSHIPStable,SHIP_NAMEhasadatatypeofVARCHAR2(20).
AllothercolumnsareNUMBER.Nowconsiderthefollowingquery(note
thatlinenumbershavebeenaddedforreadability):
HowmanyrowswilltheSELECTstatementreturn?
A.Nonebecauseofasyntaxerrorresultingfromadatatype
conflictinline4
B.Nonebecauseline5isaskingforSHIPnamesthatcontainan
underscoreafterthestring'Codd',andnonedo
C.2
D.1
8.AssumealltablenameandcolumnnamereferencesintheSQL
statementthatfollowsarevalid.Thatbeingsaid,whatiswrongwiththe
syntaxofthefollowingSQLstatement?
A.IntheWHEREclausethereisasyntaxerrorbeforetheword
CAPACITY.
B.Itneedstohaveeitheranequalsignoranot-equalsign.
C.IntheWHEREclausethereisasyntaxerrorafterthewordIN.
D.Thereisnothingwrongwiththesyntax.
9.ReviewthefollowingdatalistingfortheSHIPStable:
NowreviewthefollowingSQLstatement(linenumbersareaddedfor
readability):
WhichofthefollowingstatementsistrueaboutthisSELECT
statement?
A.Thesyntaxiscorrect.
B.Thesyntaxonlines3and4isincorrect.
C.Lines3and4havecorrectsyntaxbutcouldbereplacedwithOR
LIFEBOATSBETWEEN80AND100.
D.Line5ismissingparentheses.
UseAmpersandSubstitutiontoRestrictandSortOutputatRun
Time
10.Tolistallthecurrentlydefinedvariables,use:
A.SHOWALL
B.SHOWDEFINE
C.DEFINE
D.DEFINEALL
11.Youcanuseasubstitutionvariabletoreplace:
A.Afloating-pointvalueinaWHEREclause
B.ThenameofatableinaSELECTstatement
C.Neither
D.Both
12.Considerthefollowingtext:
Whatwillhappenwhenthisscriptisexecuted?
A.Theenduserwillbepromptedtoenteranumber.
B.ThescriptwillfailbecausevRoomNumberinthefirstlinedoes
nothaveanampersandprefix.
C.TheSELECTstatementwillfailbecausethesubstitution
variableshouldnotbeprefixedbyanampersandsinceitisalready
definedwiththeDEFINEstatement.
D.TheDEFINEstatementinline1shouldbeprecededbythe
keywordSET.
13.TopermanentlydeleteasubstitutionvariablenamedTHE_NAMEsothatit
cannolongerbeused,use:
A.UNDEFINETHE_NAME
B.SETDEFINEOFF
C.REMOVETHE_NAME
D.Youcannotdeleteasubstitutionvariable.
UsetheSQLRowLimitingClause
14.AssumeyouhaveatableITEMSthatincludesacolumnSTATUS.Whichof
thefollowingstatementsissyntacticallycorrect?(Chooseallthatapply.)
A.SELECT*FROMITEMSFETCHNEXT20%ROWSONLY;
B.SELECT*FROMITEMSFETCHNEXT20PERCENTROWS
ONLY;
C.SELECT*FROMITEMSFETCHNEXT20ROWSWITH
TIES;
D.SELECT*FROMITEMSORDERBYSTATUSFETCH
NEXT20ROWSWITHTIES;
15.Considerthefollowingstatement:
AssumeyouhaveatableITEMSwithacolumnLIST_DATE.Whatis
theresultofanattempttoexecutethestatement?
A.ItwillsorttherowsbyLIST_DATEandreturnonlythefirstfour
rows.
B.ItwillsorttherowsbyLIST_DATEandreturnonlythelastfour
rows.
C.Itwillfailwithasyntaxerrorbecauseoftheuseofanegative
numberwithOFFSET.
D.ItwillfailwithasyntaxerrorbecauseoftheuseofFIRSTand
OFFSETtogether.
SELFTESTANSWERS
SorttheRowsThatAreRetrievedbyaQuery
1. B.TheORDERBYclausewilldefaulttoascendingorderfor
SHIP_ID,butCAPACITYisexplicitlydirectedtosortindescendingorder.
A,C,andDareincorrect.TheDESCdirectiveappliesonlytothe
CAPACITYcolumnintheORDERBYclause,notbothitems.Thefactthat
theORDERBYclausereferencescolumnsthatarenotintheselectlistis
irrelevant;it’sfinetodothat.TheWHEREclauseisnotrequired;it’san
optionalclause,asistheORDERBYclause.
2. C.TheLIKEoperatorismeaninglessinORDERBY.
A,B,andDareincorrect.Thestatementwillcertainlynotexecute;itwill
failtoparsebecauseofthesyntaxerroroftheLIKEoperatorinthewrong
place.Giventhatitwon’tevenparsebecauseofsyntaxerrors,itcertainly
won’texecute.
3. AandC.ORDERBYisoptional;itisnotrequired.Itisabletosort
rowsinthetablebasedonanycriteriathataremeaningfultotherow,and
whensorted,anycolumnsmaybedisplayedfromthoserowsintheselect
list,regardlessoftheORDERBYcriteria.
BandDareincorrect.ItisuniquetotheSELECTstatementanddoesnot
appearasanoptionorotherwiseinanyotherSQLstatement.Orderingby
positionisavailabletoeachindividualORDERBYexpressionanddoesnot
dependonorrequirethesameformatfromotherORDERBYexpressions.
4. B.Ifthevaluesarecharacters,thenAislessthanZ,andifwe’re
listingrowsindescendingorder,thenthegreatervalueisshownfirst.That
meansthevalueslaterinthealphabetareshownfirst.Incomparingthe
characterstrings'Michael'and'Jackson',thestring'Michael'isgreaterand
willshowfirstinthelistingbefore'Jackson'.
A,C,andDareincorrect.Ifthevaluesarenumeric,then400islessthan
800.Thatmeansinadescendingorderlisting,wherethehighervalueis
listedfirst,800wouldbelistedbefore400.Withregardtodatedatatypes,
laterdatesaregreater,andAugust29,2010,shouldlistbeforeJune25,
2010.Finally,withregardtonumericvaluestreatedasstrings,youhaveto
thinkabouthowtheywouldappearinthedictionary—thefirstcharacteris
themostimportant,andinthiscase,the7in75indicatesthatcharacter
stringishigherthan130,soinadescendingpattern,the75wouldbelisted
before130.Ifthosevaluesweretreatedasnumericvalues,itmightbea
differentsituation.Butwe’reexplicitlydirectedtotreatthemascharacter
strings.
LimittheRowsThatAreRetrievedbyaQuery
5. B.TheSELECTwillreturnonerow,forVENDOR_ID1wherethe
CATEGORYequals'Supplier'.
A,C,andDareincorrect.Thesecondrowwillbeignoredbecauseeven
thoughthesetofexpressionswithintheINclauseincludesavaluefor
'%Partner'andusesthepercentsignwildcardcharacteratthebeginning,the
wildcardisn’tactivatedbecauseLIKEisn’tpresentintheexpression.
Therefore,thestringistreatedasaliteral.Hadtherebeenavaluein
CATEGORYof'%Partner',therowwouldhavebeenreturned.Thefailure
toincludeLIKEisnotasyntaxerrorperse;it’sjustincorrectdesignofthe
SELECTstatement.Onewaytochangethisqueryintosomethingthatis
morelikelytheintendedformwouldbethis:SELECTVENDOR_ID
FROMVENDORSWHERECATEGORYIN(‘Supplier’,‘Subcontractor’)
ORCATEGORYLIKE‘%Partner’;Thatapproachwouldproducetwo
rowsandperformthequerythatwasprobablyintended.
6. A.Thisquerywillalwaysretrievezerorows,nomatterwhatthey
looklike.Theuseofthe=NULLexpressionwithintheWHEREclause
guaranteesthatfact.NothingwilleverberetrievedbecausenoSQLengine
iscapableofconfirmingthatanyvalueisequalto“Idon’tknow.”
B,C,andDareincorrect.Ifline3hadusedISNULL,asinWHERE
STYLEISNULLORWINDOWISNULL,thentheanswerwouldhave
beenC,ortworows.Also,iftheISNULLwereusedandtheORhadbeen
ANDinstead,thenonerowwouldhavebeenreturned.Regardless,the
parenthesesarecorrecthere;youareallowedtoenclosecomplete
expressionswithinparentheseswithinaWHEREclause.
7. D.ThequeryreturnstherowwithavalueofSHIP_ID=1,andno
more.TheBETWEENrangeisinclusive,sothenumber2052ispartofthe
range.
A,B,andCareincorrect.TheLENGTHvalueisnumeric,andthesetof
expressionsinsideINarestrings.However,OracleSQLwillperforman
automaticdatatypeconversionsincethestringsallcontainnumericdata
withinthemanyway,andtheoperationwillsucceed.It’snotthebestdesign,
butitworks,andyou’llneedtobeawareofthisfortheexam.Also,line5
usestheLIKEoperatortoactivatewildcardcharacterswithinthestring,and
bothoftheavailablewildcardsareused—thesingle-characterunderscore
andthemulticharacterpercentsign.Thesecombinetoindicatethatanyrow
withaSHIP_NAMEthatstartswiththestring'Codd',followedbyatleast
onecharacter,followedbyanywherefromzerotoaninfinitenumberof
additionalcharacters,willbeaccepted.
8. D.Thereisnothingwrongwiththesyntax.
A,B,andCareincorrect.Thepairofnestedparenthesesbeforetheword
CAPACITYisavalidexpressionthatmultipliesthevalueofthe
LIFEBOATScolumnandaddsthenumber57totheendofit.Thenthe
entireresultissubtractedbywhateverthevalueforCAPACITYmightbe.
Theresultofthatexpressionwillthenbecomparedtowhateveris
containedintheseriesofexpressionsaftertheINclause.Therearetwo
expressionsthere:onemultipliestheLIFEBOATSvaluetimes20,andthe
secondaddsthevaluesofthecolumnsnamedLIFEBOATandLENGTH.
Alloftheseexpressionsaresyntacticallyvalid.
9. A.Thesyntaxiscorrect.However,therearesomeissuesinvolving
thelogic,suchastheexpressiononlines3and4,whichdon’treallydo
anything.Anynon-NULLvalueforLIFEBOATSwillbefoundwiththese
expressionsbecauseoftheORoperatoronline4.Itwouldmakemore
senseforthatoperatortobeAND,butregardless,itissyntacticallycorrect.
B,C,andDareincorrect.Lines3and4haveaccuratesyntax,buttheOR
atthebeginningofline4shouldprobablybeanAND.Sinceitisnot,then
BETWEENwouldnotbeanequivalentsubstituteheresinceBETWEEN
cantestonlyforarangeandessentiallyservesasareplacementoftheAND
combination,asinLIFEBOATS>=80ANDLIFEBOATS<=100.Line5
doesn’tneedanyparentheses.Theywouldn’thurtanythingnecessarily;
theyjustaren’trequired.
UseAmpersandSubstitutiontoRestrictandSortOutputatRun
Time
10. Ciscorrect.TheDEFINEcommandonalinebyitselfwilldisplayall
existingvariablesdefinedforthesession.
A,B,andDareincorrect.TheSHOWALLcommandlistsallsystem
variables.SHOWDEFINEwillshowthecurrentstateoftheDEFINE
systemvariable.DEFINEALLisnotavalidcommand.
11. Discorrect.Youcanuseasubstitutionvariabletoreplacejustaboutany
textinaSQLstatement,includingvaluesandSQLkeywords.
A,B,andCareincorrect.Whileyoucanuseasubstitutionvariableto
replaceafloating-pointnumberandalsothenameofatable,youcando
both,sothebestanswerisD.
12. Aiscorrect.Theenduserwillbepromptedtoenteranumberintwo
places.ThefirstpromptwillresultfromthePROMPTstatement.However,
PROMPTisnotassociatedwithACCEPT,soitwillneitherwaitfora
responsenorcaptureanyinput.However,theSQLscripthasavalid
ampersandsubstitutionvariablethatwillprompttheusertoenteravalue
forrnbr,atwhichpointtheenduserwillbeabletoprovideavalue,which
willcausetheSQLscripttosuccessfullyexecute.
B,C,andDareincorrect.TheDEFINEstatement’sreferenceto
vRoomNumberdoesnotrequireanampersand,butnordoesitassigna
valuetothevariablevRoomNumber.TheSELECTstatementwillnotfail
forthereasonprovided:theampersandiscorrectlyplaced.Thefirstline
statementdoesnotrequiretheSETstatementinfrontand,ifitwere
included,wouldresultinanerrorsinceSETDEFINEworkswiththe
keywordsON,OFF,orasinglecharactertobecomethenewsubstitution
variable.
13. Aiscorrect.UNDEFINEwillremoveasubstitutionvariablefrom
memoryforthesession.
B,C,andDareincorrect.SETDEFINEOFFtemporarilyturnsoffthe
abilitytoreferenceallsubstitutionvariables.Butitdoesnotpermanently
removeanyvariables;youcanSETDEFINEONandallexisting
substitutionvariablesareimmediatelyavailableforreferenceonceagain.
ThereisnoREMOVEstatement.Topermanentlydeleteasubstitution
variable,useUNDEFINEfollowedbythenameoftheparticular
substitutionvariableyouwanttodelete.
UsetheSQLRowLimitingClause
14. B,C,andDarecorrect.InthecaseofC,thelackofanORDERBY
rendersWITHTIESineffective,butthesyntaxisacceptablenonetheless.
Thereisnosyntaxerror.
Aisincorrect.ThepercentsigncannotbeusedinFETCH;usethe
keywordPERCENTinstead.
15. Aiscorrect.TheuseofOFFSETwithanegativenumberistreatedthe
sameasoftheOFFSETwerezero.TheSELECTwillFETCHthefirstrow
rowsofthesortedrowset.
B,C,andDareincorrect.Whileitistemptingtothinkthatanegative
offsetwouldstartattheendofthesortedlist,itdoesnot,andinsteadthe
negativeoffsetnumberistreatedasthoughitwereazero.Syntacticallythe
statementiscorrect,anditwillexecute.TheuseofFIRSTis
interchangeablewithNEXT;oneisrequiredandeitherisacceptable,the
choicedoesnotchangethefunctionalityofthestatement.
5
UsingSingle-RowFunctionstoCustomize
Output
T
CERTIFICATIONOBJECTIVES
5.01UseVariousTypesofFunctionsThatAreAvailableinSQL
5.02UseCharacter,Number,Date,andAnalytical(PERCENTILE_CONT,
STDDEV,LAG,LEAD)FunctionsinSELECTStatements
✓Two-MinuteDrill
Q&ASelfTest
hischapterlooksatthetopicofSQLfunctions.Functionsperformunique
tasksthatboostthecapabilitiesofSQL.TherearemanySQLfunctions,and
itisimportantthatyoubefamiliarwithallofthemandunderstandhowtouse
themindividuallyandinvariouscombinationsinsupportofvariousreal-world
scenarios.
Thischapterlooksexclusivelyatsingle-rowfunctions.Mostsingle-row
functionsareknownalsoasscalarfunctions.Scalarfunctionsreturnoneanswer
foreachonerowprocessed.Therearetwoothercategoriesoffunctions.You’ll
studythetopicofaggregatefunctionsandanalyticfunctionsinChapter7,when
youstudytheGROUPBYclauseoftheSELECTstatement.
CERTIFICATIONOBJECTIVE5.01
UseVariousTypesofFunctionsThatAreAvailablein
SQL
Functionshavethefollowingthreecharacteristics:
Theymayacceptincomingvalues,orparameters(notethatafew
functionstakenoparameters).
Theyincorporateparameterdataintosomesortofprocess;inother
words,theyperformsomesortoftaskontheincomingdata,suchasa
calculationorsomeotheractivity.
Theyreturnonesingleanswerasaresult.
Forexample,hereisaSQLstatementthatusesafunctioncalledINITCAP,
whichtakesasingleparameter:
Inthisexample,foreachrowthattheSELECTstatementprocesses,theSQL
built-infunctionINITCAPtakesthecolumnLASTNAMEandtransformsthe
textcontainedwithinsothatthefirstletterisdisplayedasacapitalletter(initial
capital,INITCAP,getit?)andtherestofthetextisinlowercaseletters.The
resultofthefunctionisdisplayedinplaceofthevaluefortheLASTNAME
column.ForeachonerowofdatareturnedbytheSELECTstatement,INITCAP
willreturnonevalue.SeeFigure5-1foranexampleofitsoutput.
FIGURE5-1 OutputofSELECTwithINITCAP
Asyoucansee,afunctionlikeINITCAPisn’tfoolproof—noticehowitdoes
anicetransformationonthefirsttworowsbutdoesn’tnecessarilyproducea
desirableresultinthethirdrow,wherethethirdletterLshouldstillbe
capitalized.ButINITCAPperformsitstask;it’suptousasdeveloperstoknow
wherebesttoapplyit.
WhenafunctionisusedinaSQLstatement,it’softensaidthatit’scalledor
invokedfromthestatement.
Functionscanbecalledfromanyplacethatanexpressioncanbecalled.In
otherwords,youcaninvokeafunctionfrom
ASELECTstatement’sselectlistandWHEREclause
AnINSERTstatement’slistofvalues
AnUPDATEstatement’sSETclauseandWHEREclause
ADELETEstatement’sWHEREclause
...andmore.
ThereareavarietyoffunctionsavailableinSQL.Thetwomajortypesare
built-inanduser-defined.
Built-infunctionsarethosethatarepartoftheSQLlanguage.They
areavailablewitheverystandardimplementationofSQL.
User-definedfunctionsarethosethatarecreatedbyusers.Theyare
writtenwithfeaturesthatgobeyondthecapabilitiesofSQL,using
languagessuchasPL/SQL.Theirconstructionisbeyondthescopeofthe
examandthereforethisbook.Thisbook—andtheexam—willdealonly
withbuilt-infunctions.
Thereareagreatmanybuilt-infunctions,andtheyfallintoseveral
categories.Themostcommoncategoriesarecharacter,number,anddata
functions.We’lldiscussconversionfunctionsinChapter6.
CharacterFunctions
Characterfunctionsareusedtomanipulatetext.Theycanbeusedtoperform
manyjobsonagivenstringsuchasanalyzeitslength(LENGTH),paditwith
extracharacters(RPAD,LPAD),trimoffunwantedcharacters(RTRIM,LTRIM,
TRIM),locateagivenstringwithinalargerstring(INSTR),extractasmaller
stringfromalargerstring(SUBSTR),andreplacetextwithinastring
(REPLACE).It’sevenpossibletosearchforstringsthataren’tnecessarily
spelledthesamebutthatsoundalike(SOUNDEX),andmore.Whentheseare
combined,thepossibilitiesaretheoreticallyendless.
NumberFunctions
Numberfunctionscanperformmathematicalanalysis.SQLcomeswithmany
functionsfordeterminingsine(SIN,ASIN,SINH),cosine(COS,ACOS,
COSH),andtangent(TAN,ATAN,ATAN2,TANH).Youcandetermineabsolute
value(ABS)ordeterminewhetheragivennumberispositiveornegative
(SIGN).
Afunctioncanroundoffvalues(ROUND)andotherwiseabbreviatenumbers
(TRUNC).
Numberfunctionscanbeincorporatedintoexpressions,andasyou’ve
alreadyseen,expressionsprovidesupportforstandardarithmeticoperationsof
addition(+),subtraction(–),multiplication(*),anddivision(/).Theseoperators
arenotfunctionsbutare,well,operators.Thepointhereisthatyoucancombine
theoperatorswithnumberfunctionstoproducepowerfulSQLstatements.
DateFunctions
ThesetofavailableSQLfunctionsincludespowerfulfeaturesforDATE
manipulation.Youcanobtainthecurrentdateandtime(SYSDATE,
SYSTIMESTAMP),“round”offdatestovaryingdegreesofdetail(ROUND),
andotherwiseabbreviatethem(TRUNC).Youcancalculatethedifferences
betweentwoormoredatesinmanyways.
Simplearithmeticoperatorswillhelpdeterminethedifferencesbetweentwo
datesintermsofdays,meaningthatifyousubtractonedatefromanother,the
resultinganswerwillbeanumberrepresentingthedifferenceintermsofdays.
Butwhatifyouwantsomethingelse,like—thedifferenceintermsofmonths?
(TheanswerisMONTHS_BETWEEN.)Therearefunctionsthatwillassistin
managingsuchtasks.Youcanaddorsubtractanentiremonthandaccountfor
spansoftimethatencompassyears(ADD_MONTHSwithapositivenumberto
addoranegativenumbertosubtract—we’llseethislater).
Whatifyouhaveaparticulardateandwanttoknowwhetherthatdatefalls
on,say,aSaturday?Thissortoffeaturecanbeaccomplishedwith“conversion”
functions,whichwediscussinthenextchapter.
OtherFunctions
Therearesomefunctionsthatdon’tquitefitintothecategorieslistedhere.
USERisagreatexample—it’sastandardfunctionthattakesnoparametersand
simplyreturnsthevalueshowingthenameofthecurrentuseraccount.Other
functionssupportadvancedfeatures,suchashierarchicalqueries.We’lllookat
thosefunctionsaswediscussvariousadvancedfeaturesthroughoutthebook.
CERTIFICATIONOBJECTIVE5.02
UseCharacter,Number,Date,andAnalytical
(PERCENTILE_CONT,STDDEV,LAG,LEAD)
FunctionsinSELECTStatements
ThissectionlooksindetailatmanyoftheSQLfunctions.You’lllearnhowthey
work,andyou’llseeexamplesofhowmanyofthemcanbeused.Foreach
functiondiscussedhere,weaddress
Thefunction’slistofparameters
Whethereachparameterisrequiredoroptional
Whattaskthefunctionperforms
Theoutputfromthefunction
Ratherthanpresentanalphabeticallistingofthem,I’lllisttheSQLfunctions
presentedinthissectionlogically,andI’lldescriberelatedfunctionstogether.
ThisisnotanexhaustiveanalysisofallthefunctionsavailableinSQLbutis
neverthelessrathercomprehensive,focusingprimarilyonthefunctionsthatare
mostcommonlyused.Remember,functionsareanexamobjective.It’s
importanttobefamiliarwithhowthemajorfunctionswork.Thissection
describesthemostcommonlyusedfunctionsthatarelikelytobeontheexam.
TheDUALTable
Beforewegetstarted,IneedtomentiontheDUALtable,whichissomething
thatisn’tafunctionissuepersebutishelpfultoourpurposeshere.Itisnota
“SQL”thing,norisitnecessarilya“function”thing,butit’san“Oracle”thing
youshouldknowbeforewecontinue.TheDUALtableispresentineveryOracle
database.Thisspecialtablecontainsjustonecolumn.Thecolumnisnamed
DUMMY,andithasadatatypeofVARCHAR2(1).TheDUALtablecontains
onlyonerow.ThatrowhasavalueinDUMMYof'X'.
ThepurposeofDUALissimple—tohavesomethingtorunaSELECT
statementagainstwhenyoudon’twanttoretrieveanydatainparticularbut
insteadsimplywanttorunaSELECTstatementtogetsomeothertask
accomplished,suchasobtainingtheoutputofafunction.Forexample,you’ll
soonseeinthissectionthatthere’sanOracleSQLfunctioncalledSYSDATE
thatdisplaysthecurrentdateaccordingtotheoperatingsystemoftheserveron
whichtheOracledatabaseisinstalled.IfyouwanttogetthevalueofSYSDATE
withoutwadingthroughabunchoftabledatafromyourapplication,youcan
simplyexecutethefollowingSQLstatement:
SELECTSYSDATEFROMDUAL;
Theresult:you’llgetoneresponseforthecurrentvalueofSYSDATE,since
youknowtheDUALtablehasonlyonerow.
We’llusetheDUALtablefromtimetotimethroughouttherestofthis
chapteraswegothroughexamplesofthevariousSQLfunctions.
CharacterFunctions
Thissectionlooksindetailatthemorecommonlyusedcharacterfunctions.
Characterfunctionsdonotnecessarilydealexclusivelywithcharacterdata—
someofthesefunctionshavenumericparametersandevennumericoutput,like
INSTR.Buttheoverallspiritandintentofthesefunctionsistoperformthesort
ofdataprocessingtypicallyassociatedwithtextmanipulation.
UPPERandLOWER
Syntax:UPPER(s1)
LOWER(s1)
Parameters:s1,arequiredcharacterstring.
Process:Transformss1intouppercaseletters(UPPER)orlowercaseletters
(LOWER).
Output:Characterstring.
Example:UPPERandLOWERcanbeusefulwhenyouaredoingatextsearch
andaren’tsurewhetherthedatainthetableisinuppercase,lowercase,orboth.
YoucanuseUPPERtoforceaconversionofalldatainthetableandthen
comparethatresulttoanuppercaseliteralvalue,thuseliminatinganychanceof
missingmixed-casetextinthetable.
ThepreviousquerywillreturnrowsfromtheEMPLOYEEStablewherethe
LAST_NAMEis'McGillicutty','mcgillicutty',oranyothercombinationof
upper-andlowercaselettersthatequalsthelettersinthestring.
BewarequestionsthataskyouaboutusingUPPERorLOWERwhen
bothsidesofthecomparisonoperatorarevariables.IfyourSELECT
statementendswithsomethinglikeWHERELAST_NAME=
FAMILY_NAMEandyou’reconcernedaboutcasemismatch,thenuse
UPPERorLOWER,butonbothsidesoftheequalsigns:WHERE
LOWER(LAST_NAME)=LOWER(FAMILY_NAME).Usingthe
functionononlyonesidemightactuallyincreaseyourprobabilityofa
casemismatch.Besuretousethefunctiontostandardizethecaseof
bothvaluesbeingcompared.
INITCAP
Syntax:INITCAP(s1)
Parameters:s1,arequiredcharacterstring.
Process:Transformss1intoamixed-casestring,wherethefirstletterofeach
wordiscapitalizedandeachfollowingcharacterisinlowercaseletters.
Output:Characterstring.
Thesinglequoteescapecharacteractivates,orenables,asinglequote
charactertobedisplayedasasinglequote.Withoutit,astringlike
‘O’Hearn’isinterpretedas‘O’followedbytheletters‘Hearn’that
appearoutofcontext.Thisdoesnotwork:SELECT‘O’Hearn’FROM
DUAL.Butthisdoes:SELECT‘O”Hearn’FROMDUAL.
Example:ThefollowingisaSQLstatementthatinvokesINITCAPthreetimes.
Thefirstexamplepassesinthestring'napoleon',whichistranslatedintomixed-
caseletters.Thesecondexampletakesthestring'RedO''Brien'asinput.Notice
thatweneedtoincludethesinglequoteescapecharacterwithinthestringin
orderforasinglequotecharactertoberecognized—sincethesinglequotemark
isalsothestringdelimiter,weneedtotypetwosinglequotesinsuccession,
whichishowyouinstructSQLtotreatasinglequoteasanactualcharacter
withinthestring,andnotthestringdelimiter.Withoutthis,ourstringwould
appeartobe'RedO'followedbytextthatwouldbeseenbySQLasgibberish
andwouldresultinanerrormessage.Finally,thethirdexampledoesthesame
thingwiththestring'McDonald''s',buttheresultsofINITCAParelessthan
desirable.ThefirstdinMcDonald’sisconvertedtolowercase,andthesatthe
endofMcDonald’sisconvertedtouppercasebecauseINITCAPinterpretsthes
asthestartofanewword.
CONCATand||
Syntax:CONCAT(s1,s2)
s1||s2
Parameters:s1,s2.Botharecharacterstrings;botharerequired.
Process:Concatenatess1ands2intoonestring.
Output:Characterstring.
Example:
Here’stheequivalentusingthedoubleverticalbars:
WhileCONCATtakesonlytwoparameters,thedoubleverticalbarsyntax
canberepeatedasoftenasisnecessary.Forexample,thisapproachcreatesone
string:
Notethattheresulthereisasinglestring.Here’sanotherexample:
LPAD,RPAD
Syntax:LPAD(s1,n,s2)
RPAD(s1,n,s2).
Parameters:s1(characterstring—required);n(number—required);s2(character
string—optional;s2defaultstoasingleblankspaceifomitted).
Process:Padtheleftofcharacterstrings1(LPAD)orrightofcharacterstrings1
(RPAD)withcharacterstrings2sothats1isncharacterslong.
Output:Characterstring.
Example:Takeastringliteral'ChapterOne—IAmBorn'andpadittotheright
sothattheresultingstringis40charactersinlength.
ManySQLfunctionsareusefulinandofthemselves,butmanybecomemore
usefulwhencombinedwitheachother.Forexample,here’sacombinationof
RPADandLPADwiththeconcatenationoperators,executedagainstatable
calledBOOK_CONTENTS:
Noticethisexamplealsoincludesthreeusesoftheconcatenationoperator:
OncetoappendasingleblankaftertheCHAPTER_TITLEvalue
OncetoputasingleblankinfrontofthePAGE_NUMBERvalue
OncetocombinetheoutputofRPADandLPAD
Alsonotetheuseofthecolumnalias"TableofContents",andnotethe
ORDERBYtoensurethattherowssortaccordingtoPAGE_NUMBER.
Asyouwillseeinthisbook,whichisfocusedonexampreparation,SQL
featuresanumberofpowerfulformattingcapabilities.Itisworthnoting,
however,thatreportingtoolsoftenareusedtoperformreportformattingina
professionalsettingandcanoffermorefeaturesandeaseofuse.Thesetools
maybelimitedtoparticularoperatingsystemplatforms,whereasSQL’s
formattingcapabilitiesexistwhereverSQLisavailable.
Ifthe“OntheJob”codesampleshowingLPADandRPADcombined
withstringconcatenationisamysterytoyou,thenstopandstudyit.The
exammayincludequestionsaboutseveralfunctionscombinedand
nestedwithineachother,liketheBOOK_CONTENTSsampleshownin
thesample.Wewilladdressnestedfunctionslaterinthischapter.
LTRIMandRTRIM
Syntax:LTRIM(s1,s2)
RTRIM(s1,s2)
Parameters:s1,s2—botharecharacterstrings.s1isrequired,ands2isoptional
—ifomitted,itdefaultstoasingleblankspace.
Process:Removesoccurrencesofthes2charactersfromthes1string,from
eithertheleftsideofs1(LTRIM)ortherightsideofs1(RTRIM)exclusively.
Output:Characterstring.
Notes:Idealforstrippingoutunnecessaryblanks,periods,ellipses,andsoon.
Example:Removedashesfromastring.Notethatwechoosetoincludea
columnalias'Result'torenametheoutputcolumnforthisquery.
TRIM
Syntax:TRIM(trim_infotrim_charFROMtrim_source)
Parameters:
trim_infoisoneofthesekeywords:LEADING,TRAILING,BOTH—ifomitted,
defaultstoBOTH.
trim_charisasinglecharactertobetrimmed—ifomitted,assumedtobea
blank.
trim_sourceisthesourcestring—ifomitted,theTRIMfunctionwillreturna
NULL.
Process:SameasLTRIMandRTRIM,withaslightlydifferentsyntax.
Output:Characterstring.
Example:Trimoffthedashesattheendofthestring'Seventhousand--------'.
Notewechoosetouseacolumnalias'TheOutput'torenametheoutputcolumn.
LENGTH
Syntax:LENGTH(s)
Parameters:sisthesourcestring(required).
Process:Identifiesthelengthofagivenstring.
Output:Numeric.
Example:Determinethelengthofareallylongandfamousword.
INSTR
Syntax:INSTR(s1,s2,pos,n)
Parameters:s1isthesourcestring(required);s2isthesubstring(required);pos
isthestartingpositionins1tostartlookingforoccurrencesofs2(optional,
defaultis1);nistheoccurrenceofs2tolocate(optional,defaultis1).Ifposis
negative,thesearchins1foroccurrencesofs2startsattheendofthestringand
movesbackward.
Process:Locatesastringwithinanotherstring(thusthenameofthefunction:IN
STRing).
Output:Numeric.
Example:Lookforthestring'is'within'Mississippi',startingatthefirst
characterpositionbutlookingforthesecondoccurrenceof'is'.
TheINSTRfunctionistellingusthatthesecondoccurrenceofisstartsatthe
fifthcharacterinMississippi.
SUBSTR
Syntax:SUBSTR(s,pos,len)
Parameters:s=acharacterstring,required;pos=anumber,required;len=a
number,optional.
Process:Extractsasubstringfroms,startingattheposcharacterofsand
continuingforlennumberofcharacters.Iflenisomitted,thenthesubstring
startsasposandrunsthroughtheendofs.Ifposisnegative,thenthefunction
startsattheendofthestringandmovesbackward.
Output:Characterstring.
Example:Startingwithasourcestringof'Name:MARKKENNEDY',extracta
substringoutofit,beginningattheseventhpositionandrunningtotheendof
thestring.
SOUNDEX
Syntax:SOUNDEX(s)
Parameters:s=thesourcestring,required.
Process:TranslatesasourcestringintoitsSOUNDEXcode.
Output:Characterstring.
Notes:SOUNDEXisacodingschemefortranslatingEnglishwordsintosound-
alikepatterns.AsingleSOUNDEXvalueisrelativelyworthless.Buttwo
combinedcanbesurprisinglyhelpful.Thereasonisthatsimilar-soundingwords
tendtogeneratethesameSOUNDEXpattern.
Example:TogenerateaSOUNDEXtranslationforanyword,thefirstletter
remainsunchanged.Thenextseriesoflettersaretranslatedintoanumericcode
accordingtotherulesshowninTable5-1.Translationisperformedforeach
letteruntilthreedigitsaregenerated.Ifanylettersexistbeyondthat,theyare
ignored.Forexample,thelastnameWorthingtonhasaSOUNDEXpatternof
W635:thefirstletter,W,remainsunchanged;thesecondletter,o,isignoredper
thebottomrowoftheTable5-1;thethirdletter,r,translatesinto6accordingto
Table5-1;thefourthletter,t,translatesintoa3;andthefifthletter,h,isignored.
Again,accordingtoTable5-1(whichalsotellsustoignoretheletteri),theletter
ntranslatesinto5,andnowthatwehaveoneletterWandthreenumbers635,we
aredone;theremaininglettersinWorthingtonareignored.Thus,theSOUNDEX
codeisW635.
TABLE5-1 SOUNDEXTranslationTable
So,twowordsthatsoundalike,forexample,WorthingtonandWurthinden,
whilespelleddifferently,willneverthelessgeneratethesameSOUNDEX
pattern.Here’sanexample:
NoticehowthetwodifferentwordsproducethesameSOUNDEXpattern.
Thatmeanswecandoquerieslikethis:
NoticethatSOUNDEXisusedtwiceintheWHEREclause.Usingitonce
wouldprobablybeuseless.Thisquerywillfindcustomerswiththelastnamesof
Worthington,Wurthinden,Worthan,andevenWirthen.
TheconceptandlogicofSOUNDEXwereoriginallydevelopedintheearly
twentiethcenturyforusewiththeEnglishlanguage,withahistoryof
implementationintheU.S.CensusandU.S.NationalArchives.Duringthelate
twentiethcenturyasautomatedsystemsweredeveloped,SOUNDEXbecamea
popularfeaturetoincludeincomputersoftwaresystems,andthuswefinditin
SQL.ItsrulesareformallyadministeredtodaybytheNationalArchivesand
RecordsAdministrationoftheUnitedStatesgovernment.Buttakenote:Itwas
andisdesignedforusewiththeAmericanEnglishlanguage.Itwon’tworkquite
aswellwithwordsoriginatingfromotherlanguages,whosepronunciationrules
andpracticestendtobedifferent.Forexample,thepopularVietnamesename
Nguyenispronounced“Nwen,”buttheSOUNDEXpatternsforthosetwowords
—NguyenandNwen—areratherdifferent.So,SOUNDEXisnotperfectbut
ratherusefulnonetheless.
NumericalFunctions
Thissectiondescribesfunctionsthatdealwithnumericvalues.Somenumeric
functionsareprettysimplesuchasABS,whichtakesasinglenumericvalueand
returnsitsabsolutevalue.There’salsoSQRT,whichtakesasinglenumeric
valueandreturnsitssquareroot.
Notalloftheinputparameterstonumericfunctionsarenecessarilynumeric,
buttheoverallintentofthesefunctionsistoperformnumericanalysisand
performthesortoftaskstypicallyassociatedwithnumericmanipulationand
dataprocessing.
CEIL
Syntax:CEIL(n)
Parameters:nisrequiredandisanynumericdatatype.
Process:CEILreturnsthesmallestintegerthatisgreaterthanorequalton.
FLOOR
Syntax:FLOOR(n)
Parameters:nisrequiredandisanynumericdatatype.
Process:FLOORreturnsthelargestintegerthatislessthanorequalton.
ROUND—Number
Syntax:ROUND(n,i)
Parameters:nisrequired,isanynumber,andcanincludedecimalpoints.iisan
integerandisoptional—ifomitted,itwilldefaultto0.
Process:nisroundeddependingonthevalueofi.Ifiiszero,nisroundedoffto
thenearestwholenumber,inotherwords,zerodecimalpoints.Ifiisapositive
number,nisroundedtoiplacestotherightofthedecimalpoint.Ifiisa
negativenumber,nisroundedtoiplacestotheleftofthedecimalpoint.The
number5isroundedawayfromzero.
Output:Ifiisomitted,ROUNDreturnsavalueinthesamenumericdatatypeas
n.Ifiisspecified,ROUNDreturnsadatatypeofNUMBER.
Example:Roundoff12.355143totwosignificantdigitstotherightofthe
decimal,andalsoroundoff259.99tothenearest“tens,”inotherwords,one
digittotheleftofthedecimal.
TRUNC—Number
Syntax:TRUNC(n,i)
Parameters:nisrequired,isanynumber,andcanincludedecimalpoints.iisan
integerandisoptional—ifomitted,itwilldefaultto0.
Process:TRUNC“rounds”towardzero;inotherwords,ittruncatesthenumbers.
Output:Ifiisomitted,TRUNCreturnsavalueinthesamenumericdatatypeas
n.Ifiisspecified,TRUNCreturnsadatatypeofNUMBER.
Example:UsingthesamenumberswejustusedwiththeROUNDexample,
truncatetheminstead.
REMAINDER
Syntax:REMAINDER(n1,n2)
Parameters:n1andn2arenumbers.Botharerequired.
Process:Identifiesthemultipleofn2thatisnearestton1andreturnsthe
differencebetweenthosetwovalues.
Output:Numeric.
Example:TestREMAINDERusingthreesequentialnumberssuchas9,10,and
11,andcompareeachagainstthenumber3.Sincethefirstnumber(9)isa
multipleof3,thereisnoremainder,sotheanswerwillbe0.Thesecondnumber
(10)representsonemorenumberthanthemultiple,sotheremainderis1.Notice
whathappenswiththethirdnumber(11)—thefunctiondoesn’treturna2asyou
mightexpect.Instead,itreturnsanegative1,becausethenearestintegerthat’s
divisibleby3is12,whichisclosertothe11thanthe9.Inotherwords,
REMAINDERidentifiestheclosestmultipleofn2.Ifthemultipleishigher,
REMAINDERreturnsanegativenumbertoindicatethattheclosestmultipleof
n2ishigherthann1.
MOD
Syntax:MOD(n1,n2)
Parameters:n1andn2arenumbers.Botharerequired.
Process:PerformsthesametaskasREMAINDER,exceptMODusesFLOOR
insteadofROUNDinitsequation.
Output:Numeric.
Example:GettheMODofthesamethreenumberpairswetestedwith
REMAINDER.Notetheresultsinthethirdexample—thismightbewhatyou
would’veexpectedwithREMAINDERanddidn’tget—butyoudogetitwith
MOD.
DateFunctions
ThefollowingsectionlooksatfunctionsthatworkprimarilywithDATEdata
types.
SYSDATE
Parameters:None
Process:Returnsthecurrentdateandtimeaccordingtotheoperatingsystemon
whichtheOracledatabaseserverisinstalled.Inotherwords,ifyourSQL
statementisrunningonanOracleserverinstancefromaremotelocation,then
regardlessofthelocationofyouoryourclient,SYSDATEwillreturnthedate
andtimeoftheoperatingsystemonwhichtheserverresides.Timeinformation
iscontainedwithinSYSDATEbutdoesn’tdisplaybydefault;however,itcanbe
extractedbywayoftheTO_CHARconversionfunction,whichwewillformally
discussinChapter6,butwhichweincludeinsomeoftheexamplesthroughout
thischapter.(Note:Itcanalsobealteredbychangingthe
NLS_DATE_FORMATsessionparameter.)
Output:Date.
Example:Showthecurrentdateaccordingtotheoperatingsystemwherethe
Oracleserverisinstalled.
ROUND—Date
Syntax:ROUND(d,i)
Parameters:disadate(required);iisaformatmodel(optional).
Process:disroundedofftothenearestdatevalue,atalevelofdetailspecified
byi.disrequiredandspecifiesavalueoftheDATEdatatype.iisaformat
modelandspecifiesthelevelofdetailtowhichtheDATEvaluewillberounded
—inotherwords,tothenearestday,nearesthour,nearestyear,andsoon.iis
optional;ifiisomitted,ROUNDwilldefaulttoaformatmodelthatreturnsa
valueofdroundedtothenearestwholeday.Valuesarebiasedtowardrounding
up.Forexample,whenroundingofftime,12noonroundsuptothenextday.
FormatmodelsarecoveredinChapter6whenwediscusstheTO_CHAR
conversionfunction.Butwe’llincludeoneintheexamplethatfollowstogive
youanideaofwhataformatmodelis.
Output:Date.
Example:ThisexampleshowsaSELECTstatementwiththreeexpressionsin
theselectlist.ThefirstisSYSDATE,whichreturnsthecurrentdate.Thesecond
isthesamedate,roundedtothenearestmonth,asspecifiedbythe'MM'format
model.Thethirdisthesamedateroundedtothenearestyear,asspecifiedbythe
'RR'formatmodel.
Iftheoptionalsecondparameterisomitted,theDATEvaluewillberounded
tothenearesthour.
Wehaven’tyetcoveredtheconversionfunctionTO_CHAR,butwearegoing
touseittodemonstratehowROUNDbehaveswhentheoptionalformatmodel
isomitted.Withoutaformatmodel,ROUNDwillroundadatetothenearest
wholedate.Thisroundingisperformedbutisnotnecessarilyobvioussince,by
default,valuesoftheDATEdatatypedon’tdisplaythehour.Thevalueforhour
isstoredwithinaDATEdatatypenonetheless,asarealsothevaluesforminute
andsecond,butyoumustuseaconversionfunctiontodisplaythosevalues.
That’swhereTO_CHARcomesinhandy.
Toseehowthisworks,we’llusetheROUNDfunction,theupcoming
TO_CHARfunction(explainedlaterinthischapter),andatechniqueknownas
nestedfunctions.(Wewillcovernestedfunctionsinmoredetaillaterinthis
chapter.)
Inthefollowingexample,theSELECTstatement’sfirstcolumntakesthe
valueofSYSDATEandusestheTO_CHARconversionfunctiontodisplaythe
valueofSYSDATEwithaformatmodeltodisplaythedateandthetime.
Thesecondcolumndoesthesamethingwithonechange:itusesROUNDto
roundoffthevalueofSYSDATE.NotetheimpactofROUND.
UsingtheTO_CHARconversionfunction(explainedinChapter6),wereveal
thatthecurrentsystemdateandtimeisDecember3,2016,5:35:17a.m.,which
is5:35inthemorning.ButwhenSYSDATEisrounded,wegetDecember3,
2016,12:00:00.
ThisroundingisperformedbyROUNDregardlessofwhetherTO_CHARis
used.ThepointisthatROUNDworkswhethertheresultsarevisibleornot.
ROUNDisbiasedtoroundup.Inthecaseofadatewitha12noontime,
ROUNDwillroundthedateuptothenextcalendarday.
ThefirstcolumnisAugust1and12noon.Thesecondcolumnstartswiththe
samevaluebutusesROUND.Sincethetimeofdayis12noon,theresultrounds
uptothenextwholeday:August2.
Notethatwe’llexploremoreabouthowDATEvaluesandformatmodels
workwhenwediscusstheTO_CHARandTO_DATEconversionfunctions.
TRUNC—Date
Syntax:TRUNC(d,i)
Parameters:disadate(required);iisaformatmodel(optional).
Process:PerformsthesametaskasROUNDfordates,exceptTRUNCalways
roundsdown.
Output:Date.
Example:
NEXT_DAY
Syntax:NEXT_DAY(d,c)
Parameters:disadate,required;cisatextreferencetoadayoftheweek,
required.
Process:Returnsavaliddaterepresentingthefirstoccurrenceofthecday
followingthedaterepresentedind.
Output:Date.
Example:ShowthefirstoccurrenceofaSaturdayfollowingMay31,2019.
LAST_DAY
Syntax:LAST_DAY(d)
Parameters:disadate,required.
Process:Returnsthelastdayofthemonthinwhichdfalls.
Output:Date.
Example:ShowthelastdaysofFebruaryin2020and2021.
ADD_MONTHS
Syntax:ADD_MONTHS(d,n)
Parameters:disadate,required;nisawholenumber,required.
Process:Addsnmonthstodandreturnsavaliddatevaluefortheresult.
Output:Date.
Example:AddonemonthtoJanuary31,2017,andfourmonthstoNovember1,
2017.
ThereisnoSUBTRACT_MONTHSfunction.Instead,useADD_MONTHSto
addanegativenumberofmonths,andyou’llsubtracttheminstead.
MONTHS_BETWEEN
Syntax:MONTHS_BETWEEN(d1,d2)
Parameters:d1andd2aredates,required.
Process:Determinesthenumberofmonthsbetweenthetwodates.Theresult
doesnotroundoffautomatically;iftheresultisapartialmonth,
MONTHS_BETWEENshowsarealnumberresult.Wholemonthsarecounted
accordingtothecalendarmonthsinvolved;ifthetimespans,say,aFebruarythat
has29days,thentheonemonthtimespanforthattimeperiodwillbe29days.
Inotherwords:
Notethattheanswermaybeoppositeofwhatyoumightexpect.Youmight
thinkyoucouldlistthedatesinchronologicalorderandthat
MONTHS_BETWEENwoulddeterminethedifferenceinmonths.Itwill,butas
anegativenumber.Togetapositivenumber,thefirstparameterisexpectedtobe
thegreatervalue;thesecondisexpectedtobethelesser.Eitherapproachworks,
butbesuretoconsiderthesignoftheresult;ifthesecondparameteristhe
greatervalue,theresultisanegativenumber.
Output:Number.
Example:DisplaythenumberofmonthsbetweenJune12,2014,andOctober3,
2013.
NUMTOYMINTERVAL
Syntax:NUMTOYMINTERVAL(n,interval_unit)
Parameters:n=number(required).interval_unit=oneofthefollowingvalues:
'YEAR'or'MONTH'.
Process:Convertsdateinformationinnumericformintoanintervalvalueof
time.
Output:AvalueintheINTERVALYEARTOMONTHdatatype.
Example:Thefollowingexampletakesthenumber27andtransformsitintoa
valuerepresentingatimeintervalof27months,whichequatesto2yearsand3
months,intheINTERVALYEARTOMONTHdatatype.The2-3valueshows
2yearsand3monthsistheamountoftimethatresults.
NUMTODSINTERVAL
Syntax:NUMTODSINTERVAL(n,interval_unit)
Parameters:n=number(required).interval_unit=oneofthefollowing:'DAY',
'HOUR','MINUTE',or'SECOND'.
Process:Convertsdateinformationinnumericformintoanintervalvalueof
time.
Output:AvalueofthedatatypeINTERVALDAYTOSECOND.
Example:Thefollowingexampletranslates36hoursintoitsformal
representationof1dayand12hoursinthedatatypeINTERVALDAYTO
SECOND,whichdisplaysasinglenumberforday,followedbyhours,minutes,
seconds,andfractionalseconds.
DatesandNUMBERConstants
ToperformarithmeticwithDATEvalues,usenumericliterals.Thenumberone
(1)representsoneday.Useliteralexpressionstorepresenthoursorminutes.For
example,sincethereare24hoursinaday,thenonehourcanberepresentedby
1/24.Thereare1,440minutesinaday(60minutestimes24hours),so1minute
canberepresentedby1/1440;12minutescanberepresentedby12/1440,andso
on.
Here’sanexample:
TheTIMESTAMPdatatypeworksthesamewaywiththesameresults;you
couldreplaceSYSDATEinthepreviousexamplewithSYSTIMESTAMPand
seethesameimpacttothevalues.
AnalyticalFunctions
Therearethreecategoriesoffunctionstestedbytheexam:scalar,analytical,and
aggregatefunctions.Scalarfunctionsreturnoneanswerforeveryonerow
processed.Aggregatefunctionsreturnoneanswerforasetofzerotomultiple
rows.
Analyticalfunctionsfallsomewhereinbetweenthese.Theyhavetheability
toreturnmultiplerowsfromwithinagroupofrows.BecauseagivenSELECT
statementprocessesdatarowbyrow,ananalyticfunctioncanoperateacrossa
windowofrows.Thatwindowcanbedefinedbynumbersofrowsorbylogic,
suchasatimerange.AsprocessingmovesfromrowtorowwithintheSELECT
statement’sgroupofrows,thewindowmayslide,dependingonhowtheanalytic
functionspecifiesthewindow.
Analyticfunctionsarethelastsetofoperationsperformedinaquerypriorto
theORDERBYclause,whichisthefinalprocessingstep.Forthisreason,you
cannotincludeanalyticfunctionsanywhereotherthantheSELECTlistorthe
ORDERBYclause.Inotherwords,analyticsareforbiddenintheWHERE,
HAVING,andGROUPBYclauses.
Let’slookatsomekeywordscommonlyusedwithanalyticfunctions.
OVER,PARTITIONBY,andORDERBY
ToseehowOVER,PARTITIONBY,andORDERBYareusedwithanalytics,
let’slookatanexample.First,considerthefollowingsampledataset:
WecanusetheSUMfunctionasanaggregatefunctiontoobtainthesumtotal
ofallSQ_FTvaluesfortheentiredataset,likethis:
Alternatively,wecanuseSUMasananalyticfunctionandcombineitwith
OVERtodisplayarunningtotalforeachrow,likethis(linenumbersadded):
ThecolumnRunningTotaldisplaysexactlywhatisindicated(arunningtotal
ofSQ_FTvalues)thatbuildswithinitsownwindow,whichinthiscasematches
thesetofrowsfortheSELECT.TheSubsetcolumnalsoperformstheSUM
operation,butoveraslidingwindowdefinedasthecurrentrow,theoneprior
row,andtheonefollowingrow.Forexample,theSubsetvalueforline13(Room
Number104)isthecombinationoftheSQ_FTvaluesdisplayedinlines13,14,
and15.Why?BecauseoftheseconduseoftheanalyticfunctionSUMdetailed
onlines3,4,and5.
Let’saddthePARTITIONBYoptiontoperformourSUMoversetsofrows
byvaluesfortheWINDOWcolumn.
TheinclusionofPARTITIONBYchangedthewayRunningTotalandSubset
arecalculated.Forexample,lookatthevaluesonline16.TheRunningTotal
columnrestartswithavalueof533,sincewe’vepartitionedthedataby
WINDOWandthevalueforWINDOWisnowNone(thefirstexampleofthis
valueandthereforethefirstrowinanewpartition).Therefore,therunningtotal
isthecurrentvalueforSQ_FTandnothingmore:533.
Lookatline18.TheRunningTotalvalueis2422,whichissumofthelast
rowforOcean(line15,wherethevaluewas898)andthecurrentvalueinline
18(whichis1524).Together,898plus1524is2422,thecurrentvalueforthe
runningtotalforSQ_FTforthepartitiondefinedbythoserowshavingavalueof
Ocean.
Let’schangetheORDERBYfortheSELECTbutleavetheanalytic
functionsunchanged.
NotethattheSELECTstatementre-sortstherows,butthevaluesforeach
rowremainunchanged.Forexample,online13,RoomNumber102stillshows
aSubsetvalueof2262,justasitdidinthepreviousqueryresults.Asyou’ve
alreadyseen,thevaluefortheSubsetcolumnisdeterminedbythepartitionand
orderofrowsaroundit,butthatorderwasn’tchanged;theanalyticsfunctionon
lines4through7remainsunchanged.Allthatwechangedwasthesortorderof
theSELECTstatement.
LAG,LEAD
TheLAGandLEADfunctionsaresimilartoeachother.Foragivenrowwithin
awindow,LAGshowsacolumn’svalueinthepriorrow,andLEADshowsthe
nextvalue.Forexample,considerthefollowingdataintheSHIP_CABINS
table:
Let’suseLAGandLEADinaSELECTstatementtoreturneachrow
showingthatrow’svalueforSQ_FT,thepriorrow’sSQ_FT(Lag),andthe
subsequentrow’sSQ_FT(Lead):
Inthisexample,theLAG(line2)andLEAD(line3)functionsareboth
specifiedtooperateoverawindowthatissortedbyWINDOWandSQ_FT.The
SELECTstatementusesthesameORDERBYlogic(line5).Noteline9,where
thefirstrowofoutputshowsaNULLfortheLagvalue.Thisisbecausethe
logicinline2specifieslogicresultinginthisrow(line9)asthefirstrowinthe
window(specifiedinline2).Therefore,thereisnopriorrow.
SeeFigure5-2foranillustrationoftheseconcepts.First,PARTITIONBY
separatestherowsintosets,orgroups,asspecified.Inourexample,wespecified
PARTIONBYWINDOW,andoursampledatahastwovaluesforWINDOW,so
weseparateourrowsintotwogroups—oneforaWINDOWvalueof'Ocean'
andanotherforaWINDOWvalueof'None'.Second,foreachgroup,wedefine
awindowforeachtargetrow.Inourexample,thewindowconsistsofthoserows
spanningfromonerowpriortoourtargetrowthroughtoonerowfollowingour
targetrow.Thismeansthataswestepthrougheachrow,ourwindowwillslide.
FIGURE5-2 UsingLAGandLEADanalyticalfunctionswithPARTITION
BYandOVER
Third,aswestepthrougheachrow,wewanttheSQ_FT;theLAG(SQ_FT),
whichisfromthepriorrow;andtheLEAD(SQ_FT),whichisfromthe
followingrow.Finally,wereturntheoutputandseeourresults.
Let’schangetheORDERBYlogicandseewhetherthischangestheLag
value.
TheORDERBYlogicnowsortstheoutputbyROOM_NUMBER.Butnote
theoutputinrow11:it’sthesameROW_NUMBERvalueof104,andit’sthe
sameNULLvalueforLag.ThisdemonstrateshowthewindowforLAGis
definedbyLAGitself,independentoftheORDERBYclause.
Sofarwe’velookedatLAGandLEADfunctionsthatspecificallyreturndata
fromtheonerowimmediatelypriorto(LAG)orfollowing(LEAD)thecurrent
row.Therowdifferenceisspecifiedbytheoffset,whichdefaultsto1aswe’ve
seensofarbutcanbespecifiedassomethingelse.Iftheoffsetspecifiesa
nonexistingrow,theLAGorLEADfunctionwillreturnaNULLvalue.Let’s
changetheoffsetforourLAGfunctionto2andseewhathappens.
Inthissamplelisting,seeline2fortheLAGfunctionandtheoffset
specificationof2,immediatelyaftertheSQ_FTreference.Theresultappearsin
therowsreturned.Notethatbothlines9and10showNULLvaluesforLAG.
Thisisbecausetherearenotrowsattheoffsetforthoserows.Forotherrows—
suchasline13—youseetheSQ_FTof533,andtheLagSQ_FTvalueistaken
fromtherowoffsetbytwoprior,whichinthiscaseistherowshownonline11.
STDDEV
TheSTDDEVfunctionreturnsthesamplestandarddeviationofasetofnumeric
values.Standarddeviationisamathematicalvaluerepresentingthedegreeof
distributionforagivennumericvaluewithinalargerrangeofvalues.The
standarddeviationforagivennumberisthesquarerootofitsvariance.Oracle
hasfunctionsforboth,sothevalueofagivennumber’sSTDDEVisthesquare
rootofitsVARIANCE,anaggregatefunction.UsingSTDDEV,youcananalyze
asetofnumbersanddetermineeachvalue’sdispersionfromthemean.
STDDEVisusefulwhenanalyzingasetofnumbersbylookingforcorrelations
amongthosenumbers—thosenumberswhoseSTDDEVvaluesareloware
closertothemeanthanthosenumberswithrelativelyhighervaluesofSTDDEV.
Themathematicalformulasforcalculatingstandarddeviationandvariance
aresimplebutabitinvolved.Forexample,considerthefollowingsetofvalues:
Thevaluesare1,2,4,and13,foratotaloffourvalues.
Tocalculatethevarianceandstandarddeviationofourexample,dothe
following:
Calculatetheaverage:
First,sumthevalues:1+2+4+13=20
Next,dividethesumbythetotalnumberofvalues:20/4=5.
Thisistheaverage.
Foreachnumber,determinethedifferencebetweenthenumberand
theaverage,thensquaretheresult.
Addupthesquareddifferences.Taketheresultanddividebythe
totalcountofvalueslessone:
16+9+1+64=90
90/(4–1)=90/3=30.Thisisthevariance.
Calculatethesquarerootofthevariance:
SQRT(30)=5.477.Thisisthestandarddeviation.
TheVARIANCEandSTDDEVfunctionswillperformthesecalculationsfor
you.Inthatcontext,thefunctionwillbehaveasanaggregatefunction.For
example,considerthefollowingdataset:
Nowlet’suseAVG,MEDIAN,VARIANCE,andSTDDEVasaggregate
functionsonthefullsetofrows.
ThatisanexampleofusingSTDDEVasanaggregatefunction.Inother
words,thefullsetofrowsasspecifiedbytheSELECTstatementisconsidered,
andasinglestandarddeviationiscomputedandreturned.
However,STDDEVcanalsobeusedasasingle-rowfunctionincombination
withaslidingwindowofrows.Thefollowingcodecalculatesboththe
cumulativeVARIANCEandthecumulativeSTDDEVoverthesamewindow:
Inouroutput,theVARIANCEandSTDDEVarecalculatedcumulatively,and
youcanseethefinalvaluesmatchtheaggregateresultscomputedinthe
previousexample.
PERCENTILE_CONT
ThePERCENTILE_CONTfunctionuseslinearinterpolationbetweenagiven
row’srowsetceilingandfloortocalculatethepercentilevalueforthatgiven
row.ThePERCENTILE_CONTfunctiontakesasinputapercentageyou
specify,say,.4(for40percent),anditanalyzesagroupofrowsyouspecify;then
itdeterminestheequivalentnumericvalueequaltothatpercentageofthewhole,
byusinglinearextrapolation.
Forexample,let’sdisplayrowsfromSHIP_CABINSthatshowthesquare
footageforeachcabin,butforthecabinswiththesametypeofwindow,
considerthelargestsquarefootage,calculate60percentofthatsquarefootage,
anddisplaythatamount.
HereistheformulausedforcomputingthePERCENTILE_CONT.
First,wecalculatedthepercentagebypartitioningtherowsbyWINDOW.
OursampledatahastwovaluesforWindow:OceanandNone.So,wehavetwo
groups.
Foreverygroup,PERCENTILE_CONTwillconsiderthespecified
percentage(P),whichinourcaseis.60,andthencountthenumberofrows
(RN).ForOceantherearefourrows,showninthepreviousexampleonlines10,
11,12,and15.Usingthatinformation,wecalculateatargetrow(TR)usingthe
followingformula:
TR=(1+(P*(RN−1)))
Usingournumbers,hereisthecalculation:
2.8=(1+(.6*(4–1)))
Ourtargetrownumberis2.8.WehavefourrowsforOcean,soourtargetrow
istheoreticallylocatedbetweenthesecondandthirdrows.Ourobjectiveisto
determinetheSQ_FTvaluethatwouldexistonthistheoreticalrownumber2.8.
Tocalculatethatvalue,weneedtoknowthefollowing:theCEILandFLOORof
ourtargetrownumberandthevaluesatthoserows.
Forourexample,thisinformationisasfollows:
TheCEILfor2.8is3,andthevalueforthethirdrowofOceanis
533.
TheFLOORfor2.8is2,andthevalueatthesecondrowofOceanis
205.
TheformulausedbyPERCENTILE_CONTisasfollows:
(CEIL–TR)*(ValueattheFLOOR)+(TR–FLOOR)*(ValueattheCEIL)
Inotherwords,hereisthecalculation:
(((3–2.8)*205)+((2.8–2)*533))=467.4
TheresultiswhatyouseeineachrowofOceaninouroutput.
TheintentofPERCENTILE_CONTistocomputeforagroupofrowsthe
interpolatedvalueatthespecifiedpercent(inourcase,60percent)ofthewhole
forthatgroup.
ManymorefunctionsexistinSQLthanaredescribedinthisbook.
Spacelimitationspreventmefromshowingdescriptionsandexamplesof
allofthem.Yetanyofthemanyfunctionsmayappearontheexam.Be
suretoreviewtheOracleDatabaseSQLLanguageReferenceManualand
reviewthelengthydescriptionofalltheSQLfunctionsbeforetakingthe
exam.Payparticularattentiontotheinputparametersofeachfunction,
aswellasthedatatypeofeachfunction’sreturnedvalue.
NestingFunctions
Whenafunctionisplacedwithinanexpressioninsuchawaythatitsoutput
becomestheparametervalueforanotherfunctionofthatsameexpression,the
originalfunctionissaidtobenested.Whenonefunctionisnestedwithin
another,thenestedfunctionexecutesfirst.Thenestedfunctionisalsoconsidered
theinnerfunction,asopposedtotheouterfunction,whichreceivestheoutputof
theinnerfunctionasaninputparameter.Whenfunctionsarenestedatmultiple
levels,theinnermostfunctionsexecutefirst.Youcannestmultiplefunctionsand
atmultiplelevels—soyoucannestafunctionwithinafunction,andnestthe
combinationwithinyetanotherfunction.
(Note:Onceyouincludenon-scalarfunctions(i.e.,aggregatefunctions)ina
SELECTstatement,youbegintointroducesomepotentialcomplicationsintothe
process,includingwithregardtonestedfunctions.However,thischapterlooks
exclusivelyatscalarfunctions(single-rowfunctions)sowe’lldeferconcerns
aboutnestedaggregatefunctionsuntilChapter7.
NestedFunctionsandPatterns
Here’sanexamplethatnestsonefunctionwithinanother.Thecombinationof
SUBSTRandINSTRcanbehelpfulinlocatingstringswhosepositionvaries
withinastringbutvariesrelativetoafixeddistancetoanotherstring.Oneoften
usedtechniqueishelpfulwhenworkingwithinformationaboutpostaladdresses,
whenyouarelookingforthetwo-letterabbreviationofoneoftheUnitedStates
thatisoftenfoundafterthecommaplusoneblankspace.SeeFigure5-3foran
example.ThefirstcolumnshowsthecolumnADDRESS2unchanged.Notice,
however,thateachstringcontainsatwo-letterstateabbreviationandthateach
stateisafteracommaplusoneblankspace.Wecanusethatconsistentpatternto
ouradvantage.ThesecondcolumnshowshowwecanusetheINSTRfunction
tofindtheexactlocationofthatcommaineachindividualrowofADDRESS2.
Finally,thethirdcolumnshowshowwecannesttheoutputofeachrow’s
INSTRresultwithinaSUBSTRfunction.Byadding2totheresultsofINSTR
(oneforthecomma,oneforthespace),welocatetheprecisestartofthetwo-
letterstateabbreviationwithineachoccurrenceofADDRESS2andthusareable
toextractthevalueforthethirdcolumn,STATE.
FIGURE5-3 SUBSTRandINSTRcombined
Finally,notethatweordertheoutputbythefindingsofthethirdcolumn.
Prettyinterestingthatwecansortrowsofdatabyasubstringofacolumnwhose
positionchangeswithineachrowofthetable—butit’sentirelypossiblethanks
tothepresenceofaconsistentpatterninthedataandtheuseofnestedfunctions
tolocateandworkwiththoseconsistentpatterns.
CERTIFICATIONSUMMARY
SQLfunctionsperformawidevarietyoftasks,rangingfrommathematical
calculationstotextanalysisanddateconversions.Functionscanbecalledfrom
almostanytypeofSQLstatementandfromavarietyoflocationswithinthose
SQLstatements.SQLfunctionscanbecalledfromtheSELECTstatement’s
selectlistandWHEREclause,fromtheINSERTstatement’svaluelist,fromthe
UPDATEstatement’sSETclauseandWHEREclause,andfromtheDELETE
statement’sWHEREclause.
Afunctiontakesanywherefromzerotomultipleinputparametervalues.
Eachfunctiondoessomesortofprocessingthatincorporatestheinput
parameters,andperhapssomeotherdataaswell.Eachfunctionreturnsexactly
oneresult.
Characterfunctionsperformtasksassociatedwithstringmanipulationand
textanalysis.CharacterfunctionsincludeUPPER,LOWER,INITCAP,
CONCAT,LPAD,RPAD,LENGTH,INSTR,SUBSTR,andothers.Character
functionsacceptinputparametersthatmaybecharacterdataandmayinclude
otherdatatypes,suchasnumericparameters.Andwhileeachcharacterfunction
returnsexactlyonevalue,asdoallfunctions,thecharacterfunctionsdonot
necessarilyreturncharacterdataastheirreturnvalue.Forexample,LENGTH
returnsanumberindicatingthelengthofagivencharacterstring.Buteach
performsataskassociatedwithcharacterstrings.
Numberfunctions,alsoreferredtoasnumericfunctions,performanalysison
numbers.TheyincludeROUND,REMAINDER,MOD,andothers.
Datefunctionsworkwithdateanddatetimeinformation.Datefunctions
includeSYSDATE,ROUND(fordates),TRUNC(fordates),NEXT_DAY,
LAST_DAY,ADD_MONTHS,MONTHS_BETWEEN,
NUMTOYMINTERVAL,NUMTODSINTERVAL,andothers.
Analyticalfunctionscanbeusedtoperformanalysisononeormoresetsof
rowswithinthelargerrowsetofaSELECTstatement.Thosesubsetscan
includeslidingwindowsthatvarywitheachtargetrowandcanbespecifiedby
rownumbersorbyvaluerangeusinglogic.Somefunctionscanbeusedas
aggregatestocalculateasingletotalvalueorasanalyticalfunctionswithin
subsetsofrows—likeSUM,whichcanbeusedasananalyticalfunctionto
calculatearunningtotal.
AnalyticalfunctionsincludeSTDDEV,whichcalculatesstandarddeviation,
andPERCENTILE_CONT,whichcalculatesapercentilevalueusinglinear
interpolationwithinaspecifiedsetorsubsetofrows.Otherfunctionsperform
tasksthatmayormaynotperformprocessingononeormoredatatypes.
Functionsmaybenestedwithineachothersothattheoutputofonefunction
servesastheinputparameterofanother.
✓TWO-MINUTEDRILL
UseVariousTypesofFunctionsThatAreAvailableinSQL
MostSQLfunctionsacceptoneormoreinputparameters.Afew
takenoparameters.
Eachfunctionreturnsonevalue;nomore,noless.
SQLfunctionsperformtasksofvariouskinds.
FunctionscanbeincludedanywhereaSQLexpressioncanbe
included,providedthattherulesofdatatypesarerespected.
FunctionscanbeincludedintheWHEREclauseoftheSELECT,
UPDATE,andDELETEstatements.
FunctionscanbeincludedintheSELECTexpressionlist,INSERT
valuelist,andUPDATESETclause.
UseCharacter,Number,Date,andAnalytical
(PERCENTILE_CONT,STDDEV,LAG,LEAD)Functionsin
SELECTStatements
Characterfunctionsincludetextcleanupandconversionfunctions.
UPPER,LOWER,andINITCAPcanmanagethecaseofastring.
LPADandRPADcanpadastringwithspecifiedcharacters.
INSTR,SUBSTR,CONCAT,andLENGTHcanbeusedtodivide
upandputtogetherdifferentstrings.
Numericfunctionsperformanalysisandcalculations.
TRUNCalwaysroundstowardzero.
REMAINDERandMODarevariationsondivisionandleftover
values.
PARTITIONBYspecifiessetsofrowswithintheSELECT
statementrowset.
OVERcanbeusedtospecifyaslidingwindowofrows.
TheLAGandLEADfunctionsdrawdatafrompriorand
subsequentrows.
TheanalyticalfunctionSTDDEVcalculatesstandarddeviations
cumulatively.
OtherfunctionsincludeLEASTandGREATEST.
SELFTEST
Thefollowingquestionswillhelpyoumeasureyourunderstandingofthe
materialpresentedinthischapter.Chooseonecorrectanswerforeachquestion
unlessotherwisedirected.
UseVariousTypesofFunctionsThatAreAvailableinSQL
1.Whichofthefollowingistrueoffunctions?
A.Theyneverreturnavalue.
B.Theyoftenreturnavalue.
C.Theyalwaysreturnavalue.
D.Thereisnoconsistentanswertowhethertheyreturnavalueor
not.
2.Whichofthefollowingistrueofcharacterfunctions?
A.Theyalwaysacceptcharactersasparametersandnothingelse.
B.Theyalwaysreturnacharactervalue.
C.Theyaregenerallyusedtoprocesstextdata.
D.TheygenerallyhavethelettersCHARsomewhereinthe
functionname.
3.Built-inSQLfunctions:(Choosethree.)
A.CanbeinvokedfromaDELETEstatement’sWHEREclause.
B.ArewrittenbySQLdevelopersandalsoknownas“user-
defined”functions.
C.AreavailableforusefromtheUPDATEstatement.
D.AreavailableforusewithinaSELECTstatement’sWHERE
clause,aswellastheSELECTstatement’sexpressionlist.
4.Theoutputofafunctionmaybeused:(Choosethree.)
A.Asaninputparametervaluetoanouterfunction.
B.AsacolumnofoutputinaSELECTstatement.
C.AsaninputvaluewithintheVALUESlistofanINSERT
statement.
D.AsanalternativetothekeywordSETinanUPDATEstatement.
UseCharacter,Number,Date,andAnalytical
(PERCENTILE_CONT,STDDEV,LAG,LEAD)Functionsin
SELECTStatements
5.Considerthefollowing:
SELECTMOD(5,3),REMAINDER(5,3)FROMDUAL;
Whichofthefollowingwillbetheresult?
A.1,2
B.−1,2
C.2,1
D.2,−1
6.ConsiderthefollowingSQLstatement:
SELECTSOUNDEX('Donald')FROMDUAL;
WhichofthefollowingismostlikelytobetheoutputofthisSELECT
statement?(Choosethebestanswer.)
A.D543
B.DON3
C.Donalt
D.Donalk
7.YouaretaskedtocreateaSELECTstatementtosubtractfivemonths
fromthehireddateofeachemployeeintheEMPLOYEEStable.Which
functionwillyouuse?
A.LAST_DAY
B.SUBTRACT_MONTHS
C.LAG
D.Noneoftheabove
8.ReviewthisSQLstatement:
SELECTSUBSTR('2009',1,2)||LTRIM('1124','1')FROMDUAL;
WhatwillbetheresultoftheSQLstatement?
A.2024
B.221
C.20124
D.Asyntaxerror
9.ReviewthisSQLstatement:
SELECTTRUNC(ROUND(ABS(-1.7),2))FROMDUAL;
WhatwillbetheresultoftheSQLstatement?
A.1
B.2
C.−1
D.−2
10.ReviewthisSQLstatement:
SELECTLASTNAMEFROMCUSTOMERSWHERELASTNAME=
SOUNDEX('Franklin');
Whatisapossibleresultforthequery?
A.Franklyn
B.Phrankline
C.Ellison
D.Noneoftheabove
11.ReviewthisSQLstatement:
SELECTMONTHS_BETWEEN(LAST_DAY('15-JAN-12')+1,'01-APR-
12')FROMDUAL;
Whatwillresultfromthisquery?
A.>2(somenumbergreaterthan2)
B.2
C.–2
D.<−2(somenumberlessthannegative2)
12.ThePERCENTILE_CONTfunction:
A.ReturnsthesameresultasAVG
B.ReturnsthesameresultasVARIANCE
C.CanbeusedwithROWStospecifyaslidingwindow
D.CanbeusedwithPARTITIONBYtospecifygroupsofdata
13.TheORDERBYinanOVERclause:
A.MustmatchtheORDERBYintheSELECTstatement
B.ReplacestheORDERBYintheSELECTstatement
C.OperatesindependentlyoftheORDERBYintheSELECT
statement
D.Noneoftheabove
14.TheLEADfunctionreturnsdatafrom:
A.ArowpriortothecurrentrowasspecifiedbytheLEAD
function’sORDERBYclause
B.ArowfollowingthecurrentrowasspecifiedbytheSELECT
statement’sORDERBYclause
C.TheLAGfunction’swindow’sspecifiedcolumn
D.TherowspecifiedbytheLEADfunction’soffset
15.Analyticfunctionsareprocessed:
A.AsthefirstsetofoperationspriortotheSELECTcolumnlist
processing
B.AsthefirstsetofoperationsbeforeprocessingtheWHERE
clause
C.AsthelastsetofoperationsbeforeprocessingtheWHERE
clause
D.AsthelastsetofoperationsbeforeprocessingtheORDERBY
clause
SELFTESTANSWERS
UseVariousTypesofFunctionsThatAreAvailableinSQL
1. C.Theyalwaysreturnasinglevalue.
A,B,andDareincorrect.Functionsmayormaynotreturnavalue,each
functionisdifferent,anditisnotpossibletomakeasinglestatementthat
canneverreturnavalue.
2. C.Theyaregenerallyusedtoprocesstextdata.
A,B,andDareincorrect.Theydonotallacceptcharactersasinput
parameters.Some,suchasSUBSTR,takenumericinputparameters.They
donotalwaysreturnacharactervalue—LENGTHdoesnot.Andtheydo
notallhaveCHARintheirfunctionname.
3. A,C,andD.Thefunctionsthatarereviewedinthischapterare
knownasbuilt-infunctionsandareavailabletobeusedanywhereinSQL
whereanexpressioncanbeused.ThatincludesaSELECTstatement’s
expressionlist,anUPDATEstatement’sSETclauses,anINSERT
statement’slistofinputvalues,andtheWHEREclauseofanySQL
statement—SELECT,UPDATE,andDELETE.
Bisincorrect.Built-infunctionsanduser-definedfunctionsareseparate
categories.Bothcanbeinvokedfromthesameplaces—suchasaSELECT
statement’sWHEREclause—butuser-definedfunctionsarewrittenin
languagessuchasPL/SQL.Thischapterlookedonlyatbuilt-infunctions.
4. A,B,andC.Anestedfunctionpassesitsoutputtotheouter
functionasaninputparametertothatouterfunction.Afunctioncanbeused
asacolumninaSELECTstatement,suchasSELECTSYSDATEFROM
DUAL.AfunctioncanbeuseasinputforanINSERTstatement.
Disincorrect.AfunctioncannotreplaceakeywordwithinaSQL
statement.Havingsaidthat,itisnotuncommontowriteaquerythat
producesoutputthatisitselfSQLcodeandthenexecutethosedynamically
generatedSQLstatements.Theoriginalqueryusesstringconcatenationand
otherfeaturestogenerateoutputthatisitselfaseriesofSQLstatements.
ButprocessrequiresacombinationofSQLandSQL*Plusfeaturesused
togethertoproduceoutputthatisSQLcodefreeofoutputheadings,titles,
andotherextraneouscontentsothattheoutputiscomposedofpurelySQL
statements.Theprocesscan“spool”outputtoafilethatisexecutableand
executethatfilewithaSQL*Plusstatementtoautomaticallyexecutethose
statements.ButthatapproachincludesfeaturesofSQL*Plusanditis
outsidethescopeofthisbook.
UseCharacter,Number,Date,andAnalytical
(PERCENTILE_CONT,STDDEV,LAG,LEAD)Functionsin
SELECTStatements
5. D.MODdividesthefirstnumberbythesecondandreturnsthe
remainderrelativetothefirstnumber.REMAINDERdoesthesamething
butreturnstheremainderrelativetothesecondnumber. A,B,andCare
incorrect.Aisincorrectforbothfunctions.Bwouldbecorrectifthe
functionswerereversed.CiscorrectforMODbutisincorrectfor
REMAINDER.
6. A.TheoutputofSOUNDEXisafour-charactercodestartingwith
thesameletteroftheinputvalue,followedbyathree-digitcodebasedon
theSOUNDEXlogic.
B,C,andDareincorrect.DON3cannotbecorrectbecausethesecond,
third,andfourthcharactersshouldbedigits.CandDareincorrect—they
arebothsound-alikepossibilitiesthatmighthavethesameSOUNDEX
code,buttheythemselveswillnotbetheoutputofSOUNDEX.
7. D.Noneoftheabove.ThefunctionyouwantisADD_MONTHSso
thatyoucanaddnegativefivemonthstothecurrentdate.
A,B,andCareincorrect.TheLAST_DAYfunctionreturnsthelastday
ofagivenmonth.ThereisnoSUBTRACT_MONTHSfunction.LAGisan
analyticfunctionthatshowsacolumn’svalueinthepriorrow.
8. A.TheSUBSTRfunctionwillextractdatafromthestring('2009')
startingatthefirstposition(1)andlastingfor2characters(2),resultingin
ananswerof'20'.TheLTRIMfunctionwilltrimoffalloccurrencesofthe1
fromtheleftsideof'1124',resultingin24.Thetworesultsareconcatenated
togetherforafinalresultof2024.
B,C,andDareincorrect.Thereisnosyntaxerrorpresentinthecode.
Eventhoughthevaluesevaluatedbybothfunctionsincludedatathat
consistsofnumerals,bothareenclosedinsinglequotesandthereforeare
treatedascharacterdata.Eveniftheywerenotenclosedinsinglequotes,
SQLwouldperformanautomaticdatatypeconversionanyway.
9. A.Theresultwillbe1.TheABSfunctiondeterminesabsolute
value.Astheinnermostfunction,itwillbeevaluatedfirst,resultinginan
answerof1.7.TheROUNDfunctionwillprocessthevalueof1.7and
roundittothenearesttwodigitstotherightofthedecimalpoint,andthe
resultwillstillbe1.7.Finally,TRUNCwilltruncatethevaluedowntoa
one.
B,C,andDareincorrect.Werethe“,2”omittedasthesecondargument
intheROUNDfunction,optionBwouldbecorrect,whichistosaythe
SQLstatementwouldreturnavalueof2.IfboththeABSandROUND
functionswereeditedoutandonlyTRUNCremained,thestatementwould
return−1,makingCthecorrectanswer.WereonlytheABSfunction
removedfromthestatement,thestatementwouldreturn−2,meaningthatD
wouldbethecorrectanswer.
10. D.Noneoftheaboveiscorrect.
A,B,andCareincorrect.SOUNDEXshouldbeappliedtoboththe
sourceandcomparisontext,butinthisSELECTstatementitisusedon
onlyonesideofthecomparisonoperator.Specifically,theSOUNDEX
codeforthetextstring'Franklin'istheliteralvalue'F652'.Istheliteral
'F652'equaltothevalueforLASTNAMEintheCUSTOMERStable?
Probablynot,unlesssomeonewithanamelikeFreddyF652happenstobe
acustomer.ThepurposeofSOUNDEXistoconvertthevaluesthemselves
(inthiscaseLASTNAMEvalues)usingSOUNDEXandthencompare
thoseconvertedvaluestoagivenSOUNDEXcodetofindsound-alikes.So
ifyouarelookingforlastnamesthatareequaltoorsoundlikeFranklin,
youwouldwanttoconvertthelastnamesandthencomparethose
convertedlastnamestotheSOUNDEXcode,likethis:
WHERESOUNDEX(LASTNAME)=‘F652’
orlikethis:
WHERESOUNDEX(LASTNAME)=SOUNDEX(‘Franklin’)
TouseSOUNDEXeffectively,textonbothsidesofthecomparison
operatorneedtobeSOUNDEXvalues.
11. C.Theanswerwillbe–2.First,theLAST_DAYfunctionwilltransform
thevalueof'15-JAN-12'to'31-JAN-12',andthentheresultofthatwillbe
addedto1,sothatthefirstofFebruarywillresult:'01-FEB-12'.The
differencebetweenthatdateand'01-APR-12'willbeanegative2.
A,B,andDareincorrect.Ifthe“+1”wereremoved,thentheanswer
wouldbelessthan−2.Ifthedateswerereversed,theanswerwouldbe
greaterthanpositive2.Thereisnoscenariowhereanobviousandminor
editwouldresultinavalueofpositive2.
12. D.PERCENTILE_CONTcanbeusedwithPARTITIONBYtospecify
groupsofdatawithinasingleSELECTstatement.
A,B,andCareincorrect.PERCENTILE_CONTisnotthesameas
AVG.Itusesamorecomplexsetoflogictointerpolateananswer.Itdoes
notreturnthesameresultasVARIANCE.PERCENTILE_CONT,unlike
someotheranalyticalfunctions,doesnotsupportslidingwindows.
13. C.TheORDERBYinanOVERclauseoperatesindependentlyofthe
ORDERBYintheSELECTstatement.
A,B,andDareincorrect.TheORDERBYintheOVERclausedoesnot
havetomatchtheORDERBYintheSELECT,nordoesitreplacethe
ORDERBYintheSELECT.Thetwoexistindependentlysothatyoucan
logicallysortasubsetinsupportofthecalculationswithinanalytics,
particularlytherow-by-rowcalculations.
14. D.TheLEADfunctionreturnsdatafromtherowspecifiedbytheLEAD
function’soffset.Theoffsetdefaultsto1butmaybespecifiedwithinthe
LEADfunction.
A,B,andCareincorrect.LEADreturnsdatafromarowfollowingthe
currentrow,asdeterminedbytheoffsetandsortedbytheLEADfunction’s
ORDERBYclause.ThewindowforLEADisspecifiedbyLEAD,not
otheranalyticfunctions,whichspecifytheirownwindows.
15. D.Analyticfunctionsarethelastsetofoperationsperformedbeforethe
ORDERBY.Forthisreason,theymustbespecifiedintheSELECTlistor
ORDERBYonly,nottheWHERE,GROUPBY,orHAVINGclauses.
A,B,andCareincorrect.Theseoptionsareexclusivetothecorrect
answer,whichisthattheanalyticfunctionsareperformedatthelastsetof
operationspriortotheORDERBY.
6
UsingConversionFunctionsandConditional
Expressions
T
CERTIFICATIONOBJECTIVES
6.01DescribeVariousTypesofConversionFunctions
6.02UsetheTO_CHAR,TO_NUMBER,andTO_DATEConversionFunctions
6.03ApplyGeneralFunctionsandConditionalExpressionsinaSELECT
Statement
✓Two-MinuteDrill
Q&ASelfTest
hischapterlooksatthetopicsofconversionfunctionsandconditional
expressions.Conversionfunctionsperformdatatypetransformationfromone
generaltypesuchasnumerictoanothersuchastext.Conditionalexpressions
performevaluationsandreturndatausingcriteriayouspecifyinthecontextof
SQLstatements.These(andother)SQLfunctionscanbeusedinvarious
portionsofaSELECTstatement,includingthecolumnlist,theWHEREand
ORDERBYclauses,withinJOINandotherexpressions,andmore.
CERTIFICATIONOBJECTIVE6.01
DescribeVariousTypesofConversionFunctions
Thepurposeofconversionfunctionsistoconvertdatavaluesfromonedatatype
toanother.Conversionfunctionsdon’tchangevalues;theychangethevalue’s
datatype.Forexample,youmighthavefinancialvaluesstoredwithintext
containingdollarsignsandothersymbolsthatamathematicalfunctionmight
rejectwithanerrormessage.Butyoucoulduseaconversionfunctiontoconvert
thefinancialdatawithinthetexttoanumericdatatype,andthenusewhatever
numericfunctionsyouneedontheconvertedvalues.
InChapter2wediscusseddifferentdatatypes.Mostdatatypesfallintoone
ofthreegeneralcategories:numeric(NUMBER),text(VARCHAR2),anddates
and/ortimes(DATE,TIMESTAMP,etc.).InChapter5welookedatseveral
functionsthatperformvariousoperationsondata,butmanyrequirethat
incomingdatabeofaparticulardatatype.Forexample,mathfunctions
generallyoperateonnumericdatatypeinputs.TheADD_MONTHSfunction
usesDATEvaluesasinput.
Conversionfunctionscanbeusedtotransformdatafromonetypetoanother.
Thiscanenabletheconverteddatatobecomeinputtosomeotherfunction
requiringtheconverteddatatype.Forexample,atablewithacolumnNOTE
thatisdeclaredVARCHAR2(30)mightcontainavalueof'1230'.Thathappens
tobeanumber.Youcouldperformadatatypeconversiononthisvaluetofirst
convertitsdatatypefromtexttonumeric.Onceconverted,itcanbeusedas
numericinputtoafunctionthatcanacceptonlyincomingnumericdata.
Alternatively,iftheconversionattemptfails,thatfailuremightindicatethevalue
isnotnumericafterall;perhapsthezeroattheendof'1230'isreallytheletterO.
Yourbusinessrulesmayrequiresuchaconfirmation.Thisishelpfulwhenyou
areworkinginrealtimewithdata,particularlyifyouarecodingSQLtobeused
onarecurringbasiswithinanapplication.
Conversionfunctionscandomorethanchangeavalue’sdatatype.Some
conversionfunctionsalsofeatureformattingcapabilities.Forexample,ifyou
haveaDATEdatatypethatcontainsrawinformationabouthours,minutes,
seconds,andthedateonthecalendar,youcantransformitintoatextstringthat
spellsouttheentiredateindetail,suchas“Thursday,JulytheFourth,Seventeen
Seventy-Six.”(Seethe“OnTheJob”afterthischapter’sTO_DATEsectionfor
theSQLtomakethatwork.)Similarly,conversionfunctionscantransformraw
numericdataintofinancialformatsthatincludedollarsignsandsymbolsfor
otherinternationalcurrency,properplacementforcommasandperiods
accordingtovariousinternationalformats,andmore.
ThenextsectionwilldelveintothespecificsofTO_NUMBER,TO_CHAR,
andTO_DATE,aswellassomeotherconversionfunctions.However,beforewe
gettothat,wewouldberemissifwedidnotalsoaddressthetopicofexplicit
andimplicitconversion—let’saddressthatnow.
ExplicitandImplicitConversion
Theuseofadatatypeconversionfunctionisexplicitconversion.Butthereis
alsoanimplicitconversionfeatureinSQL.Itoccursautomaticallyin
circumstanceswhereSQLwilldetermineonitsownthatadatatypeconversion
isrequiredbutnotspecified.Whenthishappens,SQLwillperformthedatatype
conversionautomatically.Forexample,let’ssayyouwanttoconcatenatea
numericvaluewithatextstring.
NotethatthesecondvalueintheSELECTlistisnumericyetisincludedina
textfunctionforperformingstringconcatenation.SQLwillrecognizetheintent
andwillperformthecharacterconversionfunctionautomatically,executingthe
SQLstatementappropriately.
SQLisgenerouswithsuchautomaticconversions.Thischaracteristicisa
greatconveniencewhenperformingquick,adhocinteractionswiththedatabase.
Here’sanotherexample(linenumbersadded):
Inthisexample,theSUBSTRfunctioninline3returnsastringvalueof3,
whichisautomaticallyconvertedtoanumericdatatypesinceanumericiswhat
isrequiredfortheADD_MONTHSfunctioninlines2and3.
However,whenyourobjectiveistoprogramcodeforreuse,itisconsidered
gooddesigntoavoiddependenceonautomaticdatatypeconversionandcode
yourintentionsexplicitly.Andmorethansimplygooddesignandcommon
practice,itisalsotheformalrecommendationofOracleCorporationtouse
explicitconversionwherepossible.Thepracticeensuresyourintentinyourcode
isclear,documented,andsupportedinthefuturewhenthedatabaseconditions
maychangeforvariousreasons.Forexample,youmightcodeanimplicitdata
typeconversionthatassumesaparticulartable’scolumnwillhappentocontain
numericvalueswithinaVARCHAR2column.Butwhatif,inthefuture,those
valueschangetoalphabeticallyspelledreferencessuchas“two”and“three”
insteadof“2”and“3”?Theerrormessagethatwillresultwillreferencethe
formulainwhichtheautomaticconversionwasattempted.Ifyouweretohave
originallyusedanexplicitdatatypeconversionfunction,thentheerrormessage
wouldindicatethefailureoccurredattheattempttoconvertthedatatype,
makingthetrueproblemclearandresultinginafasterandmorestraightforward
troubleshootingeffort.
Whencomparingtextandnumericvalues,suchasinanIFstatement,thetext
valuewillgenerallybeconvertedtonumeric.Butnotalways:
Notethatthenumberthreeisnotgreaterthanthenumbertwenty.Butthefirst
SELECTcomparesthetwovaluesastextvalues,andinthedictionary,yes,the
textvalue'3'comesbeforethetextvalue'20',wherethefirstcharacterdefines
theprimarysortingcriteria.Whenbothvaluesaretreatedasnumericvalues,
obviously20isalargernumber.Whenthetwovaluesarecomparedsothatonly
oneisnumeric,anautomaticdatatypeconversionisperformedonthetext
value,andsuddenlybothvaluesaretreatedasnumeric.
Thefollowingaresomeadditionalrulesforautomaticdatatypeconversion:
Numericvalueswillgenerallynotconvertautomaticallytodates.
Numericandtextdatawillgenerallynotconvertautomaticallyto
verylargesizedtypes,suchasLOB,CLOB,RAW,andothers.
Verylargetypeswillgenerallynotautomaticallyconverttoeach
other,suchasfromBLOBtoCLOB,forexample.Somewill,butmany
won’t.
Explicitdatatypeconversionisclear,iseasiertoread,andresultsinbetter
performance;implicitconversionisnotasefficient.Forexample,anindex
canspeedtheperformanceofaSELECTstatementbasedonhowtheWHERE
clauseisphrased,butifthatWHEREclauseusessyntaxthatdependsonan
implicitconversion,theindexmightnotbeusedormightevenbeused
incorrectly.Thisdoesn'taffecttheresultingdata,ofcourse,butitmightresult
inperformancedegradation,whichresultsinlongerprocessingtime.
Additionally,Oracle’sdocumentationwarnsthatthebehaviorofimplicit
conversionissubjecttochangeacrosssoftwarereleases.Thebottomlineis
thatcertaintypesofdatatypeconversionsareperformedautomatically.But
don’tdependontheminyourcode.Useexplicitconversion,aswereviewedin
thissection.
CERTIFICATIONOBJECTIVE6.02
UsetheTO_CHAR,TO_NUMBER,andTO_DATE
ConversionFunctions
Conversionfunctionsconvertthedatatypeofanexpressionfromonedatatype
toanother.Someconversionfunctionswillalsotransformtheformatofthedata
atthesametime.
Forexample,thefollowingisanINSERTstatementthatattemptstostoredata
intotwocolumns.TheCALL_IDcolumnisoftheNUMBERdatatype.The
CALL_DATE_TZcolumnisofthedatatypeTIMESTAMPWITHTIME
ZONE.Here’sanattempttoinsertdataintothattable:
Nowlet’strythatsameINSERTstatementwithaconversionfunction.
Inthisexample,theTO_TIMESTAMP_TZconversionfunctionisusedto
sendthesamedataweusedinourpreviousINSERT.Thisparticularconversion
functionusesa“formatmodel”thatdescribestheformatofthedatatothe
database.Theformatmodelusedhereis'DD-MON-RRHH24:MI:SS'.This
helpstoensurethattheinputdataisrecognizedcorrectly.
Thenextsectiondescribesmanyconversionfunctionswithexamplesoftheir
use.
ConversionFunctions
ThemostcommonlyusedconversionfunctionsareTO_NUMBER,TO_CHAR,
andTO_DATE.Let’slookateachofthesethreefunctions.
TO_NUMBER
Syntax:TO_NUMBER(e1,format_model,nls_parms)
Parameters:e1isanexpression(required);format_modelistheoptionalformat
model.SeeTable6-1foracompletelistofelementsthatmakeuptheformat
model.
TABLE6-1 NumberFormatElements
ThereisanoptionalthirdparameterrepresentingNLSsettings.Itallowsyou
toidentifyanyofthethreeNLSparametersdefinedinTable6-2.Ifincluded,the
thirdparameterforTO_NUMBERconsistsofasinglestringthatencompasses
anyoneormoreofthosethreeNLSparameters.Forexample,thefollowingis
oneexampleofthenls_parmsparameterthatprovidesaspecificationoftwoof
theNLSparameters:
TABLE6-2 TheNLSParameters
Notethatsincethevaluesareenclosedinsinglequotesyetincludesingle
quotesthemselves,eachoccurrenceofthesinglequoteswithinthestringmust
beprecededbytheescapecharacter—whichisasinglequote—toclarifythatthe
valueisinfactasinglequoteaspartofthestring,ratherthantheendofthe
overallstringliteralvalue.
ThesevaluescanbeusedtodeclarenonstandardNLSparametervalues
withintheincominge1parameter.
Process:Transforme1fromanexpression,perhapsacharacterstring,into
anumericvalue,usingformat_modeltodeterminewhatformate1maytake
andwheretoextractthenumericvaluesfromamongtheformatting
information.
Output:Numeric.
Example:Intheexamplethatfollows,ourstartingvalueisastring,
'$17,000.23'.Thisisn’tanumericdatatypebutacharacterstringcontaininga
dollarsignandacomma.Theformatmodelhereexplainsthatthedollarsign
isasymbolandmakesitclearwherethesignificantnumericdatacanbe
foundinthesourcecolumn.The9elementinthefollowingexampleisnota
literalnumber9butratheranelementoftheformatmodelthatindicatesthe
presenceofanydigit.Itisrepeatedtoindicatetheupperboundofacceptable
values.Finally,theoutputisdisplayed—arawnumericvalueextractedfrom
thecharacterstring'$17,000.23'.
Hereisasimilarexampleshowingtheuseofthenls_parmsparameter:
Inthisexample,theincomingvalueshowsadecimalpointtomark
“thousands”andthecommatomarkthedecimalpoint.Thenls_parmsvalue
clarifiesthistotheTO_NUMBERfunction,alongwiththeformatmask,andthe
incomingvalueisinterpretedandtranslated,asshowninthedisplayedoutput.
SeeTable6-1foracompletelistoftheelementsthatcanbeincludedina
numericformatmodel.
TO_CHAR
TheTO_CHARfunctionconvertsdatafromvariousdatatypestocharacterdata.
ThereareactuallythreedifferentTO_CHARfunctions.Theyare,inthemost
technicalofterms,threeoverloadedfunctions.Anoverloadedfunctionisone
thatsharesanamewithoneormorefunctions,butwhereeachisdifferentiated
bytheirrespectiveparameterlists.Eachparameterlistspecifiesadifferent
function,eventhoughtheymighthavethesamename.
TherearethreeversionsofTO_CHAR:onewhosefirstparameterisa
characterstring,anotherwhosefirstparameterisadate,andanotherwhosefirst
parameterisnumeric.
ThefollowingsectionsdescribeeachofthethreeTO_CHARfunctions.
TO_CHAR—CHARACTER
Syntax:TO_CHAR(c)
Parameters:ciseitheranNCHAR,anNVARCHAR2,aCLOB,oran
NCLOB.
Process:TransformstheincomingparameterintoaVARCHAR2.
Output:VARCHAR2.
Example:
Therearesituationswhereyou’llworkwithdatatypesthatcannotaccept,for
example,CLOBdatabutcanaccepttheoutputofTO_CHAR,suchasa
VARCHAR2datatype.
TO_CHAR—NUMBER
Syntax:TO_CHAR(n,format_model,nls_parms)
Parameters:nisanumber(required).Theparameterformat_modelis
optional.Aformatmodelconsistsofoneormoreformatelements,whichyou
sawearlierlistedinTable6-1.Thenls_parmsvalueisthesameparameter
yousawearlierwiththeTO_NUMBERfunction.
Process:Transformsnintoacharacterstring,usingtheoptionalformat
modelforguidanceastohowtoformattheoutputwithanyspecialcharacters
thatmaybedesired,suchasdollarsignsorotherfinancialsymbols,special
handlingofnegativenumbers,andsoon.
Output:Character.
Example:Formatthenumber198withadollarsignandpenny
specification.
TO_CHAR—DATE
Syntax:TO_CHAR(d,format_model,nls_parms)
Parameters:disadateoradateinterval(required).Theparameter
format_modelisoptionalandcanbeusedtoformatdatainavarietyofways.
SeeTable6-3fordetailsonformatmodelsfordatedatatypes.Thenls_parms
parameteristhesameasyousawearlierfortheTO_NUMBERfunction.
TABLE6-3 DateFormatElements
Output:Character.
Example:Here’sanexampleoftheuseofadateformatmodel,asdescribedin
Table6-3:
Changingtheformatmaskstomixedcasesendsanimpliedmessagetomix-
casetheoutputaswell.
Addingthethindicatorintroducesanadditionalimprovement.Theinclusion
ofthwillappendwhateverisappropriateafterthedate—for1,you’llget“1st”;
for2,you’llget“2nd”;andsoon.Here’sanexample:
TheformatmodelisthesecrettoextractingthetimevaluesfromSYSDATE.
Here’sanexample:
NoticeinthisexamplewecanuseeitherAMorPMtoindicatewherewe
wantthemorning/afternoonindicatortobelocatedandwhetherwewantitto
includeperiods.WhetherweuseAMorPMmakesnodifference;the
appropriateindicatorwillappearwherevertheformatmodeldirects,asshownin
theprecedingexample.
TheSYSDATEfunctiondisplaysthecurrentdatebydefaultasstoredonthe
serveronwhichthedatabaseisinstalled.Butburiedinsideofitisalsothetime
ofday,inhours,minutes,andseconds.Thefullsetofdatacanbeextractedfrom
SYSDATEwiththeformatmodelparametersoftheTO_CHARfunction,as
showninTable6-3.Butbeware,thereisdangerhere,andit’syetanother
exampleofhowtrickySQLcanbe.TakealookatthisSQLstatement:
Seeanythingwrongwithit?Perhapsnot.Mostdevelopersdon’t;thiscantrip
upeventhemostexperiencedandseasonedofSQLprofessionals.Tryitonany
databaseinstance,anditwillwork,andtheoutputwillprobablyappeartobe
correct.Butlookcloselyatthevaluedisplayedforthatportionoftheformat
modelrepresentedbyMM.ThenlookatTable6-3.MMisnotminutes;itis
months.Ifyouwantminutes,youneedtouseMI,asinHH:MI:SS.Watchthis
one,folks;it’stricky—thesyntaxistechnicallycorrect,theexecutionwillbe
successful,andthetestdatalookscorrectataglance.Butit’sstillwrong.The
sharpeyeofacertifiedOracleDatabaseSQLExpertshouldflagthis.
TheYYandRRformatmasksinterprettwo-digityearrepresentations
differently.Seethefollowingquery:
Notethatadateenteredas50isinterpretedas2050bytheYYYYformat
mask,but1950isenteredasRRRR.SeeTable6-3formoreinformationabout
YYYYandRRRR.
Thehourofmidnightisthestartingmomentofanewday.Thevaluefor
midnightisstoredconsistentlybutrepresenteddifferentlybasedontheformat
modelused.Ina24-hourformat(HH24),midnightisrepresentedas00,butina
12-hourformat(HH12orHH)midnightisrepresentedas12.Forexample,we
canspecifyadateliteralat12midnightasastring,convertittoaDATEdata
typeusingTO_DATE,andthenuseTO_CHARwithaformatmasktoillustrate
thepoint.
Thefirstcolumntakesthedateliteral'07-JUL-16'atmidnight,convertsittoa
DATEdatatypein24-hourformat,andthenconvertsitbackto24-hourformat.
ThesecondcolumnstartswiththesameDATEvaluebutusesTO_CHARto
displaythedatevaluewitha12-hourformat.RememberthatthePMformat
modelspecifiesthedisplayofPMorAM,whicheverisappropriate.
The24-hourformatshowsmidnightas00,andthe12-hourformatshows
midnightas12.Thevaluestoredisnotdifferent.Onlytherepresentationis
different,inaccordancewiththeformatmodelspecified.
TO_DATE
Syntax:TO_DATE(c,format_model,nls_parms)
Parameters:c=acharacterstring(required);format_modelisaformat
modelaccordingtoTable6-3.Thenls_parmsvalueisthesameparameter
yousawearlierwiththeTO_NUMBERfunction.
Process:TransformthevaluecontainedwithincintoavalidDATEdata
typebystructuringformat_modeltodescribehowthecharacterstringis
formed,identifyingthedateinformationaccordingly.
Output:Date.
Example:Convertanonstandarddaterepresentationtothedefaultformat.
Itiscommontonestconversionfunctionsforvariousreasons—forexample,
toleveragetheirformattingcapabilities.Forexample,todeterminethe
weekdayofaparticulardate,youcannestTO_DATEwithinTO_CHAR,like
this:SELECTTO_CHAR(TO_DATE('04-JUL-1776'),'Day')FROMDUAL,
orfortheformattedversion:SELECTTO_CHAR(TO_DATE('04-JUL-1776'),
'fmDay,Month"the"Ddthsp,Year')FROMDUAL.
AdditionalConversionFunctions
WehavejustdiscussedtheTO_NUMBER,TO_CHAR,andTO_DATE
functions,themostwidelyknownoftheconversionfunctions,eachofwhichis
calledoutasanobjectiveontheexam.Thissectionaddressesadditional
conversionfunctionsthathavealsobeenknowntoappearonthecertification
examandareincreasinglyimportantinworkingwithSQLdatatypessuchas
TIMESTAMP,intervals,andothers.
TO_TIMESTAMP
Syntax:TO_TIMESTAMP(c,format_model,nls_parms)
Parameters:cisacharacterdatatype(required);format_modelmustdefine
theformatofccorrespondingtoTIMESTAMPformatmodelelements
(optional).ThedefaultrequirementisthatcmustbeintheTIMESTAMP
format.Thenls_parmsvalueisthesameparameteryousawearlierwiththe
TO_NUMBERfunction.
Process:ConvertscdatatotheTIMESTAMPdatatype,whichdiffersfrom
DATEinthatitincludesfractionalseconds.Theformat_modelparameter
definesthepatternofc’sdateinformationtothefunctionsothevarious
elementsofTIMESTAMPareidentified—informationforyear,month,day,
hours,minutes,seconds,andfractionalseconds.
Output:AvalueintheTIMESTAMPdatatype.
Example:Hereisacharacterrepresentationofadate.Theformatmodelis
includedtodefinethepatternandinformtheTIMESTAMPfunctionwhere
theDDinformationis,wheretheMONinformationis,andsoon.
TO_TIMESTAMP_TZ
Syntax:TO_TIMESTAMP_TZ(c,format_model,nls_parms)
Parameters:cisacharacterstring(required).Theformat_modelvaluemust
definetheformatofccorrespondingtoTIMESTAMPWITHTIMEZONE
formatmodelelements(optional).Thedefaultrequirementisthatcmustbein
theTIMESTAMPformat.Theoptionalnls_parmsvalueisthesameparameter
yousawearlierwiththeTO_NUMBERfunction.
Process:TransformscintoavalueofTIMESTAMPWITHTIMEZONE,where
format_modeldefinestheformatinwhichcstorestheTIMESTAMPWITH
TIMEZONEinformation.Thetimezonewilldefaulttothatdefinedbythe
SESSIONparameter.
Output:AvalueintheTIMESTAMPWITHTIMEZONEdatatype.
Example:Convertthecharacterstring'17-04-201616:45:30'toadatatypeof
TIMESTAMPWITHTIMEZONEbyprovidingaformatmask.
Notethatthereisn’taconversionfunctionthatspecificallyconvertsvalues
intotheTIMESTAMPWITHLOCALTIMEZONEdatatype.Forthat,use
CAST,describedlaterinthischapter.
TO_YMINTERVAL
Syntax:TO_YMINTERVAL('y-m')
Parameters:yandmarenumberscontainedwithinastring(required).
Process:Transformsyandmintotheyearsandmonthsinaformatofthe
datatypeINTERVALYEARTOMONTHS.
Output:AvalueintheINTERVALYEARTOMONTHSdatatype.
Example:Convertthecharacterexpressionshowingfouryearsandsix
monthsintothedatatypeINTERVALYEARTOMONTHS.
TO_DSINTERVAL
Syntax:TO_DSINTERVAL(sql_format,nls_parms)
Parameters:sql_formatisacharacterstringintheformatrequiredforthe
INTERVALDAYTOSECONDdatatype,whichis'DAYSHH24:MI:SS.FF'.
Forexample,'1514:05:10.001'istheINTERVALDAYTOSECOND
representationfor15days,14hours,5minutes,and10.001seconds.The
nls_parmsvalueisthesameparameteryousawearlierwiththe
TO_NUMBERfunction.
Process:Transformstheincomingvaluerepresentedinsql_formattoa
valueoftheINTERVALDAYTOSECONDdatatype.
Output:AvalueintheINTERVALDAYTOSECONDdatatype.
Example:Thefollowingconvertsavaluerepresenting40days,8hours,30
minutes,and0.03225secondsintotheINTERVALDAYTOSECONDdata
type:
NUMTOYMINTERVAL
Syntax:NUMTOYMINTERVAL(n,interval_unit)
Parameters:n=number(required).interval_unit=oneofthefollowing
values:'YEAR'or'MONTH'.
Process:Convertsdateinformationinnumericformintoanintervalvalue
oftime.
Output:AvalueintheINTERVALYEARTOMONTHdatatype.
Example:Thefollowingexampletakesthenumber27andtransformsit
intoavaluerepresentingatimeintervalof27months,whichequatesto2
yearsand3months,intheINTERVALYEARTOMONTHdatatype.The2-
3valueshowsthat2yearsand3monthsistheamountoftimethatresults.
NUMTODSINTERVAL
Syntax:NUMTODSINTERVAL(n,interval_unit)
Parameters:n=number(required).interval_unit=oneofthefollowing:
'DAY','HOUR','MINUTE',or'SECOND'.
Process:Convertsdateinformationinnumericformintoanintervalvalue
oftime.
Output:AvalueofthedatatypeINTERVALDAYTOSECOND.
Example:Thefollowingexampletranslates36hoursintoitsformal
representationof1dayand12hoursinthedatatypeINTERVALDAYTO
SECOND,whichdisplaysasinglenumberforday,followedbyhours,
minutes,seconds,andfractionalseconds.
CAST
Syntax:CAST(eASd)
Parameters:eisanexpression;disadatatype.
Process:Convertsetod.Thisisparticularlyusefulforconvertingtext
representationsofdatetimeinformationintodatetimeformats,particularly
TIMESTAMPWITHLOCALTIMEZONE.
Output:Avalueintheddatatype.
Example:Inthefollowing,weconvertavalueinthedefaulttimestamp
format,presentedasaliteralvalue:
Ifwewanttouseaformatmaskforanyreason,wecannestacallto,for
example,theTO_TIMESTAMPconversionfunction,asfollows:
CERTIFICATIONOBJECTIVE6.03
ApplyGeneralFunctionsandConditional
ExpressionsinaSELECTStatement
Thepowerofconditionalexpressionsiswellknownineverysoftware
programminglanguage,mostofwhichfeatureanIFstatementwithsome
combinationofELSE,THEN,orsomethingcomparable.Conditional
expressionsevaluatedataatthetimeofcodeexecutionandchangethecode’s
logicalapproachtoprocessingbasedoncertainvaluesofdataatruntime.For
example,processingmaydeterminefirstifaparticularemployeeisinacertain
department,andifso,applyabonus;ifnot,processingmayassignapayraise.
SQLdoesnothaveanIFstatement,butitdoesfeatureconditional
expressions.TheCASEstatementevaluatesanexpressionand,basedonits
value,evaluatesoneofaseriesofexpressionsasspecifiedintheCASE
statement.TheDECODEstatementissimilar;itistheoriginalIFTHENELSE
ofOracleSQL.TheNVLfunctionisusefultoreturningrowsofdatathatmight
containNULLvalueswhereyouwanttheNULLtobereplacedwithsomething
morerelevanttoyourimmediatepurposes,perhapsanumericzero,forexample.
NULLIFisusefulforcomparingtwovaluesandreturningaNULLvaluewhen
bothvaluesareidentical.Forexample,thisishelpfulincertainsituationswhere
youaretryingtoflagonlyvariationsordiscrepanciesbetweentwodatasets.
Thissectionwilllookatthesefunctionsandexpressions.
CASE
Syntax:CASEexpression1WHENcondition1THENresult1WHEN
condition2THENresult2...ELSEresultfinalEND
Parameters:expression1canbeacolumninaSELECTstatement’sselect
listoranyothervalidexpression(required).Ifexpression1evaluatestoa
valuethatisequaltocondition1,thenthefunctionreturnsresult1.Additional
WHEN/THENcomparisonpairsmaybeincluded.Thefirstpairisrequired;
additionalpairsareoptional.AnoptionalELSEattheendwillreturnthe
valueofresultfinalifnoWHEN/THENcomparisonpairmatched.
Process:Compareallthepairstodeterminewhichvaluewillbereturned.If
novaluesmatch,resultfinalisreturned.IfnovaluesmatchandnoELSE
clauseisincluded,NULLisreturned.
Example:
ThefunctionstartswiththekeywordCASEandendswiththekeywordEND.
TheCASEexpressionmayincludeacolumnname,likethis:
NotethatinthisexampletheCASEfunctiontakesinanumericvalueand
returnsatextstring.
DECODE
Syntax:DECODE(e,search_expression,d)
Parameters:e,search_expression,anddareallexpressions.Thefirsttwo
arerequired;thethirdisoptional.
Process:eisarequiredexpression;search_expressionisaseriesofpairsof
expressions,se1andse2,eachseparatedbycommas;ifeequalsse1,then
DECODEshouldreturnse2.Otherwise,itshouldreturnd.Ifdisomitted,
DECODEwillreturnNULL.
InDECODE,twoNULLvaluesareconsideredtobeequivalent.NULL
comparedtoNULLwillproduceaTRUEresultandsendthecorresponding
valuebackifrequired.Inotherwords,theuseofNULLwithinDECODEis
interpretedasacomparisonofISNULLasopposedto=NULL.Rememberthat
=NULLwillalwaysreturnafalsesinceNULLisunknown;thinkofitas“I
don’tknow.”So,issomeparticularvalue,say,4,equalto“Idon’tknow”?Who
knows?Nobodyknows,andSQLassumesFALSEineverysuchcase.ButIS
NULLisaskingadifferentquestion:isthevalueinquestionaNULL,or
unknown,value?That’showDECODEusesNULL,inthecontextoftheIS
NULLcomparison.Inthiscase,acomparisonofaNULLvaluewillresultina
truecondition.
Output:Ifthedatatypesofeandthefirstoccurrenceofse1arecharacter,
DECODEwillreturnavalueofdatatypeVARCHAR2.Ifthedatatypesofe
andthefirstoccurrenceofse1arenumeric,DECODEwillreturnavalueof
numericdatatype.
Example:Intheexamplethatfollows,weselectrowsfromADDRESSES
bylookingattheSTATEcolumnvalueasisandthenuseDECODEto
translatethevaluesinSTATEaccordingtothesearch_expressionin
DECODE,whichinthiscaselooksatonlytwostatevaluesbutcouldhave
easilybeenexpandedtotranslate,ordecode,allofthestatevalues.Thefinal
iteminthisDECODEexampleis'Other',whichisassignedtoallvaluesof
STATEthataren’tfoundinoursearch_expressionlist—includingtheNULL
valueforSTATE.
TheDECODEfunctionisoftenreferredtoastheIF-THEN-ELSEofOracle
SQL.
NVL
Syntax:NVL(e1,e2)
Parameters:e1ande2areexpressions,botharerequired,andbothshould
beofthesamedatatype,butautomaticdatatypeconversionapplieshere,so
valuesmaybedifferentaslongastheyarecapableofbeingconvertedtothe
samedatatypeautomatically.
Process:Ife1hasavalueofNULL,thenNVLreturnsthevaluefore2.
Otherwise,itreturnse1.TheintentofNVListouseitinaquerywhere
multiplerowsarebeingreturnedandyouexpectthatperhapssomeofthe
rowsmightbeNULL.There’snothingwrongwiththatinandofitself,but
whatifyouareperformingsomesortofprocessingthatcan’ttakeavalueof
NULL?Forexample,asingleNULLvaluewithinamathematicalcalculation
willautomaticallymaketheanswerNULL.YoucanuseNVLtosubstitute
somethingmeaningfulintheplaceofNULL—suchasazero—tosatisfythe
outerfunction.
Output:Ife1hasacharacterdatatype,theoutputwillbeVARCHAR2.If
e1isnumeric,theoutputwillbenumeric.IftheoutputisNULL,thenit’s
NULL.
Example:NotethatwehavethreeexpressionsintheSELECTlistthat
follows.Thefirstsimplyshowsthatwe’reusingNVLtoreplacetheliteral
valueforNULLwithazero.Thisisuselessbyitself,butitprovesthepointof
whattheNVLfunctiondoes.Thesecondexpressionshowsanequationin
whichweadd14toNULLandsubtract4fromtheresult.Butwhatis14plus
NULL?It’sNULL.SoisNULLminus4.(Remember,NULListheabsence
ofinformation.Whatis14plus“Idon’tknow”?Theansweris“Idon’t
know,”whichisNULL.)
Sointhethirdexpression,weuseNVLtoreplaceNULLwitha0,andweget
ananswerof10.
ThepurposeoftheprecedingexampleistoshowwhatyoucandowithNVL.
Amorelikelyscenariowouldbesomethinglikethis:
ThisSQLcodeaddsthesquarefeetofaship’scabinwiththesquarefeetof
itsbalcony.Butwhatifthereisn’tabalconyandaNULLvalueisreturnedfor
BALCONY_SQ_FT?TheentireresultwouldbeNULL,unlessweusetheNVL
functionaswedointheprecedingexample.
NULLIF
Syntax:NULLIF(e1,e2)
Parameters:e1ande2arebothexpressions(required).Theymustbethe
samedatatype.
Process:Ife1ande2arethesame,NULLIFreturnsNULL.Otherwise,it
returnse1.
Output:Anexpressionmatchingthedatatypesoftheinputparameters.
Example:NULLIFisgoodforcomparingmultiplerowswhereinanolder
andnewerversionofaparticularvalueexistandyouwanttoculloutthose
thatarestillnotupdatedorperhapshavealreadybeenupdated.
Intheprecedingexample,thecolumnUPDATED_TEST_SCORErepresents
asetofvaluesthatincludesolderTEST_SCOREvaluesandthosethathave
beenrevisedforsomereason.TheNULLIFfunctionhelpsfilteroutonlythose
valuesthatrepresentchangestotheolderoriginalvalues,asevidencedinthe
thirdSELECTcolumnwiththecolumnaliasREVISION_ONLY.
CERTIFICATIONSUMMARY
Conversionfunctionstransformvaluesfromonedatatypetoanother.Themost
commonconversionfunctionsaredate,text,andnumericconversionfunctions.
Theuseofaconversionfunctiontoperformdatatypeconversionisconsidered
“explicitconversion,”asopposedto“implicitconversion”whichmayoccur
automaticallywhenyourcodeomitsanyuseofaconversionfunction,yetthe
situationrequiressuchatransformationandSQLlanguageprocessingperforms
theconversionanyway.Goodprogrammingpracticeistouseexplicitdatatype
conversiontoclarifytheintentofyourcodeandincreasetheoddsofcatching
potentialerrorsincontext,whichcanpotentiallyimprovetroubleshooting.
Certainconversionfunctionsdomorethantransformvaluesfromonedata
typetoanother.Someareusedtoapplyformatmodelstoformatvaluesfor
variouspurposes—suchastoquicklygenerateameaningfulreport,ortoextract
detailsnotpresentinthestandarddefaultpresentationofcertaintypesofdata.
Dateconversionfunctionscanbeusedtodeterminethedayoftheweekfora
givendate,ortobetransformedintoatextualpresentationofthedate;numeric
conversionfunctionscanbeformattedwithfinancialformatsinaccordancewith
agivenlocale,withcommas,periods,andotherspecialcharactersas
appropriate.
ThemostcommonlyusedconversionfunctionsareTO_NUMBER,
TO_CHAR,andTO_DATE.
DatetimeconversionfunctionsincludeTO_TIMESTAMP,
TO_DSINTERVAL,andTO_YMINTERVAL.
TheCASEandDECODEfunctionsarelikeIF-THEN-ELSElogicforaSQL
statement.TheycanbeusedasanyotherfunctioninvirtuallyanyDML
statement,subquery,SELECTlist,WHEREclause,oranywhereaSQL
expressioncanbeinvoked.
CASEcomparesasingleexpressiontoaseriesofWHENconditions.Ifthe
expressionevaluatestoavalueequaltoaWHENcondition,theWHEN
condition’scorrespondingTHENresultisreturned.AnoptionalELSEresult
maybeincludedaswell.TheCASEstatementcompleteswiththekeyword
END.ThecompleteCASEstatementisconsideredanexpressionandassuch
maybeincludedanywhereinaSQLstatementthatanexpressionispermitted.
DECODEconsidersasingleexpression.Ifthatexpressionevaluatestoany
oneofaseriesofspecifiedcomparisonexpressions,thatspecifiedexpression’s
evaluatedvalueisreturned.Anoptionalfinalexpressionmaybeincludedasthe
defaultvaluetobereturnedintheeventthatnoneofthecomparisonexpressions
matchestheevaluatedexpression.
TheNVLfunctiontakestwoparametersthatarebothexpressions.Ifthefirst
expressionevaluatestoanon-NULLvalue,itisreturnedasis,butifitevaluates
toNULL,thesecondexpressionisreturned.ThepurposeoftheNVLfunctionis
toreplaceanyNULLvalueswithsomethingyouspecify.Forexample,you
mightqueryatableofnamesandwanttoreplaceanymissingmiddlenames
withtheexpression"NMN"(fornomiddlename).NVLisidealforqueriesin
whichsomevaluesmightreturnaNULLbutyoupreferinsteadtoreturnan
alternativevalue.Onecommonuseisinmathequations.Forexample,withina
mathexpression,youcanuseNVLtosubstituteanyvaluesthatmightbeNULL
withsomethingthatisn’tNULL—likeazero—toenablethemathequationsto
calculatesomethingmeaningful.
NULLIFcomparesasubjectvaluewithacandidatevalueandreturnsNULL
ifbothvaluesareidentical.Thisisusefulinreportsanalyzingdiscrepancies.
✓TWO-MINUTEDRILL
DescribeVariousTypesofConversionFunctions
Mostdatatypesfallintothecategoriesofnumeric,text,anddata
types.
ConversionfunctionsincludeTO_NUMBER,TO_CHAR,and
TO_DATE.
Ifnodatatypeconversionfunctionisexplicitlyspecified,
sometimesSQLwillperformanimplicitdatatypeconversionifit
determinesthatoneisrequired.
Implicitdatatypeconversioncanresultinunpredictableoutcomes.
Therefore,Oracleformallyadvisestheuseofexplicitdatatype
conversionswherepossible.
SQLprovidesexplicitdatatypeconversionfunctionsforeachdata
type.
UsetheTO_CHAR,TO_NUMBER,andTO_DATEConversion
Functions
TO_CHARcanconvertfromcharacter,date,ornumericdataand
intocharacterdata.
TheconversionfunctionTO_TIMESTAMPcanconverttothe
TIMESTAMPdatatype,whichisthesameasDATEbutaddsfractional
seconds.
ThefunctionsTO_DSINTERVALandTO_YMINTERVALconvert
tointervaldatatypesINTERVALDAYTOSECONDandINTERVAL
YEARTOMONTH.
ApplyGeneralFunctionsandConditionalExpressionsina
SELECTStatement
CASEandDECODEacceptoptionaldefaultvaluestoreturnifthe
evaluatedexpressiondoesn’tmatchanyofthespecifiedpossible
matches.
NVLconsidersanexpression—suchasacolumn—andevaluatesits
valuetodeterminewhetherthevalueisNULL.
NVLispopularinmathequationswherethesourcedatamightbe
NULLbuttheequationrequiressomevalue,perhapsazero.
Theexpression(SALARY+BONUS)willreturnNULLifthe
valueforSALARYisequalto,say,100,butBONUSisNULL.The
expression(SALARY+NVL(BONUS,0))willreturn100ifthe
SALARYis100andtheBONUSisNULL.
NULLIFreturnsaNULLifbothparametersareequal.
SELFTEST
Thefollowingquestionswillhelpyoumeasureyourunderstandingofthe
materialpresentedinthischapter.Chooseonecorrectanswerforeachquestion
unlessotherwisedirected.
DescribeVariousTypesofConversionFunctions
1.Conversionfunctions:
A.Changeacolumn’sdatatypesothatfuturedatastoredinthe
tablewillbepreservedintheconverteddatatype.
B.Changeavalue’sdatatypeinanequationtotellSQLtotreatthe
valueasthatspecifieddatatype.
C.AresimilartoALTERTABLE…MODIFYstatements.
D.ArenotrequiredbecauseSQLperformsautomaticdatatype
conversionwherenecessary.
2.Whichofthefollowingstatementsaretrue?(Choosetwo.)
A.Youcanuseadatatypeconversionfunctiontoformatnumeric
datatodisplaywithdollarsignsandcommas.
B.Thepresenceofanexplicitdatatypeconversiondocumentsyour
intentinthecode.
C.Dependingonthevalues,youcansuccessfullyuseanexplicit
datatypeconversiontotransformnumericvaluestotextbutnotthe
otherwayaround;youcan’texplicitlyconverttexttonumeric.
D.Animplicitdatatypeconversionperformsfasterthananexplicit
datatypeconversion.
3.Conversionfunctionscannotbeusedto:
A.Formatdatevalues
B.Convertcolumnstonewdatatypes
C.Transformdata
D.Createuser-defineddatatypes
UsetheTO_CHAR,TO_NUMBER,andTO_DATEConversion
Functions
4.Ifyouwanttodisplayanumericvaluewithdollarsignsandcommas,
whichofthefollowingisthebestapproachtotake?
A.TheTO_NUMBERfunctionwithaformatmodel
B.TheTO_CHARfunctionwithaformatmodel
C.Acombinationofstringliteralsthatcontaincommasanddollar
signs,alongwiththeCONCATfunction
D.TheMONEYdatatype
5.WhichofthefollowingSQLstatementswilldisplaythecurrenttime,
inhours,minutes,andseconds,asdeterminedbytheoperatingsystemon
whichthedatabaseserverresides?
A.SELECTTO_CHAR(SYSDATE)FROMDUAL;
B.SELECTTO_CHAR(SYSDATE,'HR:MI:SE')FROMDUAL;
C.SELECTTO_CHAR(SYSDATE,'HH:MI:SS')FROMDUAL;
D.SELECTTO_CHAR(SYSDATE,'HH:MM:SS')FROMDUAL;
6.WhichqueryreturnsanexpressionofthedatatypeINTERVALYEAR
TOMONTHSrepresentinganintervalof1yearand3months?
A.SELECTTO_YMINTERVAL('01:03')FROMDUAL;
B.SELECTTO_YMINTERVAL('01-03')FROMDUAL;
C.SELECTTO_INTERVALYM('01:03')FROMDUAL;
D.SELECTTO_INTERVALYM('01-03')FROMDUAL;
7.Youneedtodeterminethedayoftheweekforaparticulardateinthe
future.Whichfunctionwillrevealthisinformation?
A.TO_CHAR
B.DAY_OF_WEEK
C.TO_DATE
D.Noneoftheabove
8.Youaretaskedtocreateareportthatdisplaysthehoursandminutesof
thecurrentdateinareport.Whichofthefollowingwillsatisfythis
requirement?
A.TO_DATE(SYSDATE,'HH:MM')
B.TO_DATE(SYSDATE,'HH:MI')
C.TO_CHAR(SYSDATE,'HH:MM')
D.TO_CHAR(SYSDATE,'HH:MI')
9.Whichformatmaskreturnsthelocalcurrencysymbol?
A.L
B.C
C.$
D.Noneoftheabove
ApplyGeneralFunctionsandConditionalExpressionsina
SELECTStatement
10.ThepurposeofNULLIFisto:
A.ReturnaNULLifasinglecolumnisNULL
B.ReturnaNULLifasingleexpressionisNULL
C.Bothoftheabove
D.Noneoftheabove
11.Considerthefollowingquery,itsoutput,andasubsequentquery:
Whatistrueofthefinalqueryshownpreviously?
A.Itwillreturn“norowsfound”becausethereisnoPRICEof10.
B.ItwillreturnonlytherowwhereLINE_ITEMis210.
C.ItwillreturnnorowsbecausethereisnoPRICEof10.
D.Itwillreturnthreerows,butitwillnotchangethepriceforline
items100and184.
12.ConsiderthefollowingtablelistingfromthetableALARM_HISTORY:
Youaretaskedtocalculatetheaveragenumberofalarmincidentsper
dayinALARM_HISTORY.Youknowthefollowingqueryissyntactically
correct:
However,youareawarethatthevalueforINCIDENTSmightbe
NULL,andyouwanttheAVGreturnedtobecalculatedacrosseveryday
inALARM_HISTORY,notjustthenon-NULLdays.Whichofthe
followingquerieswillachievethisgoal?
13.WhichofthefollowingcanbesaidoftheCASEstatement?
A.Itconvertstexttouppercase.
B.ItusesthekeywordIF.
C.ItusesthekeywordTHEN.
D.ItsENDkeywordisoptional.
14.Considerthefollowingstatement:
WhichofthefollowingstatementsistrueofthepreviousSELECT
statement?
A.Thestatementwillfailwithacompilationerrorbecausethereis
nocolumnaliasontheNVLexpression(line1).
B.Thestatementwillfailbecauseofsyntaxerrorsonlines2and3.
C.ThestatementwillfailbecauseofthekeywordENDonthe
fourthline.
D.Thestatementwillexecutesuccessfully.
15.TheDECODEexpressionalwaysendswith:
A.ThekeywordEND
B.Adefaultexpressiontoreturnifnoothervaluematchedthe
sourceexpression
C.Bothoftheabove
D.Neitheroftheabove
SELFTESTANSWERS
DescribeVariousTypesofConversionFunctions
1. B.AconversionfunctiontellsSQLtotreataspecifiedvalueasa
specifieddatatypeforanysubsequentusewithinaparticularexpression.
A,C,andDareincorrect.Conversionfunctionsdonotchangeanything
aboutatable’sstructure.TheyarenotthesameasALTERTABLE...
MODIFY,whichcanbeusedtochangeacolumn’sdatatypeinatable.
Conversionfunctionshaveatemporaryimpact;theirscopeistheSQL
expressioninwhichtheyarebeingused,andtheyleavenolastingimpact
beyondthat.However,theyareusedsinceyoucannotcountonSQL’s
implicitdatatypeconversionworkingasyourbusinessrulesmightrequire
atalltimes.ItistheformalrecommendationofOracletouseexplicitdata
typeconversionwhererelevantandrequiredandtonotcountonimplicit
conversionwithinyourproductionSQLcode.
2. AandB.Oneofthemostcommonusesofconversionfunctionsis
toformatdata,suchasnumericanddateformatting.Theuseofanexplicit
datatypeconversionfunctionclarifiesyourpurposeinthecodeandyour
expectationaboutwhatsortofvaluesarecontainedwithinthevarious
elementsofyourcode.
CandDareincorrect.Youcanexplicitlyconvertnumerictotextand
alsotexttonumeric;ifthevaluesmakesense,thenconversionwillsucceed.
Inotherwords,youcanconvertthetextvalue'1'tothenumericvalueof1.
Also,explicitconversionfunctionsperformmoreefficientlythanareliance
onimplicitconversions.
3. BandD.Conversionfunctionscannotchangethedatatypeofa
columntosomethingelse.Theycannotcreateuser-defineddatatypes.
AandCareincorrect.Conversionfunctionsarecapableofformatting
andtransformingdate,text,andnumericdataastheyperformdatatype
conversions.
UsetheTO_CHAR,TO_NUMBER,andTO_DATEConversion
Functions
4. B.TheTO_CHARfunctionwouldwork,alongwithaformat
model,suchasTO_CHAR(rawNumber,'$999,999.99').
A,C,andDareincorrect.TheTO_NUMBERfunctionworkswith
formatmasks,butitconvertsfromcharacterstonumericvalues,notthe
otherwayaround.Youmaybeabletouseacombinationofconcatenation
andstringliterals,butitwouldbepainstakinglydifficult,andpeoplewould
laughatyou.Suchanapproachwouldbeparticularlyproblematicina
dynamicenvironmentwherethesignificantnumbersinvolvedcould
fluctuate.ThereisnoMONEYdatatypeinOracleSQL.
5. C.Thecorrectformatmaskis'HH:MI:SS'.
A,B,andDareincorrect.TO_CHARwithnoformatmaskexecutes
successfullybutdoesnothingandshowsthedatealone.ThereisnoSE
formatmask.Districky—itworksandproducesoutput,buttheMM
formatmaskindicatesmonths,notminutes,andislogicallyincorrect.
6. B.TheTO_YMINTERVALfunctioniscorrect,withthesingle
parameterofastringcontainingtwonumbers,separatedbyadash,where
thefirstrepresentsyearsintheinterval,andthesecondrepresentsthe
numberofmonthsintheinterval.
A,C,andDareincorrect.ThereisnoTO_INTERVALYMfunction.The
useofacolonisinappropriateintheTO_YMINTERVALfunction.
7. A.Here’sanexample:SELECTTO_CHAR(SYSDATE,'Day')
FROMDUAL;.
B,C,andDareincorrect.ThereisnoDAY_OF_WEEKfunction.The
TO_DATEfunctionconvertstexttodateformatsbutdoesn’tincludethis
sortofformatting.
8. D.TheTO_CHARfunctionformatsdates.TheHHandMIformat
masksdisplayhoursandminutes,respectively.
A,B,andCareincorrect.TheTO_DATEfunctionwillnotaccept
SYSDATEasincomingdata.TheTO_DATEfunctiondoesperform
formatting,butgenerallynotfordisplayinareport;instead,itcanconvert
formatstotransformincomingdataintotheformatrequiredforstoragein
thedatabase.TheTO_CHARfunctioniscorrect,buttheMMformatmask
isusedformonths,notminutes.
9. A.Listhelocalcurrencysymbolformatmask.
B,C,andDareincorrect.Cisnotaformatmask.The$specifiesthatthe
USDcurrencysymbolshouldbeusedbutwilloverrideanyconflicting
localcurrencysymbol.
ApplyGeneralFunctionsandConditionalExpressionsina
SELECTStatement
10. D.ThepurposeofNULLIFistotaketwoparameters,bothofwhichare
expressions,andreturnaNULLvalueifthetwoexpressionsevaluateto
NULL.Onecommonuseistoanalyzetwosetsofdataandlookfor
discrepancies;theinclusionofNULLIFcanomitthoseexpressionsthatare
identicalinordertohighlightthediscrepancies.
A,B,andCareincorrect.NULLIFdoesnotparticularlycarewhether
oneexpressionisNULLornot;itassesseswhethertwoexpressionsare
identical.
11. D.TheSELECTwillreturnallthreerows,changingonlythepriceforline
items210to10,andnothingelse.
A,B,andCareincorrect.NVLisnotusedheretosearchforrowsthat
matchthespecifiednumber,whichinthiscaseis10.Itisusedtotransform
anyoccurrencesofNULLvaluestothespecifiednumber,which,again,in
thiscaseis10.NVLisnotincorporatedintotheWHEREclauseinthis
situation,soitisonlyfilteringdatawithintherows,notfilteringtherows
themselves;therefore,itisnotgoingtolimittherowsreturned,asisthe
casewithallfunctions.NVLisnodifferentinthatregard.
12. B.TheNVLfunctionshouldbeappliedfirst,whichmeansitshouldbethe
innermostnestedfunction.NVLrequirestwoparameters. A,C,andD
areincorrect.NVLdoesnotworkwithoneparameter,itrequirestwo.The
NVLfunctionshouldbetheinnermostfunctioninordertofirsttransform
NULL
13. C.ItusesthekeywordTHEN.Here’sanexample:SELECTCASE
CAPACITYWHEN2074THEN'MEDIUM'WHEN4000THEN
'LARGE'ENDFROMSHIPS;.
A,B,andDareincorrect.Itdoesnotconverttext;CASEisaconditional
expression.ItdoesnotincorporatetheIFkeywordbutisoftensaidtobe
sortofanIF-THEN-ELSEexpression.ThekeywordENDisrequired.
14. B.TheSQLstatementcontainsthesamesyntaxerrorrepeatedtwice:the
keywordWHEREisincorrect.ThekeywordWHENiscorrect.
A,C,andDareincorrect.Thecolumnaliasisoptionalbutnotrequired.
ThekeywordENDiscorrectandrequiredforCASE.Thestatementwill
notexecutesuccessfullyduetotheWHEREsyntaxconflict.
15. D.Neitheransweriscorrect.ThekeywordENDisusedwithCASEbut
neverappearswithDECODE.DECODEdoesacceptadefaultvalue,but
itsinclusionisoptionalandnotrequired.
A,B,andCareincorrect.
7
ReportingAggregatedDataUsingtheGroup
Functions
I
CERTIFICATIONOBJECTIVES
7.01DescribetheUseofGroupFunctions
7.02GroupDatabyUsingtheGROUPBYClause
7.03IncludeorExcludeGroupedRowsbyUsingtheHAVINGClause
✓Two-MinuteDrill
Q&ASelfTest
nChapter5,welookedatsingle-rowfunctions,whichprocessrows
individuallyandreturnoneanswerforeachrow.Thischapterreviews
functionsandfeaturesinSQLthathavetheabilitytoprocesszeroormorerows
asalogicalset,analyzedtogethertoreturnasingle-rowresult.Thesefunctions
arereferredtoasgroupfunctionssincetheyacceptinputfromgroupsofzeroor
morerows.WewillalsolookattwonewclausesintheSELECTstatement:the
GROUPBYclauseanditscompanion,theHAVINGclause.GROUPBYand
HAVINGareusedwithgroupfunctions.Youwillneedtofullyunderstandthese
conceptsandtheirusesinordertopasstheexam.
CERTIFICATIONOBJECTIVE7.01
DescribetheUseofGroupFunctions
ThefunctionswereviewedinChapter5arereferredtoassingle-rowfunctions.
Thetermsingle-rowmeansthateachfunctionreturnsonevalueforeachonerow
itencounters.Anothertermforsingle-rowfunctionisscalarfunction.
Thefocusofthischapteristhecategoryreferredtoasgroupfunctions.A
groupfunctionreturnsonevalueforeachsetofzeroormorerowsitencounters.
Therearetwokindsofgroupfunctions:aggregateandanalyticfunctions.
Aggregatefunctionsarealsoknownasmultirowfunctions.Aggregate
functionsaretypicallyusedwithaSELECTstatementthatselectsmanyrows,
wheretheaggregatefunctionscansasetofrowsandreturnsasingleanswerfor
allofthem.
Analyticfunctionscanconsiderasetofrowsandprocessthemasasingleset
orasvaryingsubsets,includingoverlappingsubsets.Inotherwords,asingle
sourcerowmayberepresentedinmorethanonesubsettedrowsofananalytic
function’soutput.
You’vealreadyseenthesyntaxofanalyticfunctionsinChapter5whenwe
lookedattheSTDDEV,PERCENTILE_CONT,LAG,andLEAD.We’llseethat
syntaxagainintheupcomingsectionabouttheRANKandDENSE_RANK
groupfunctions.
Manyofthegroupfunctionscanbeusedasanaggregatefunctionandalsoas
ananalyticfunction,basedonthesyntaxoftheparticularstatementinwhichitis
invoked.Thereareafewexceptions.Table7-1summarizessomeofthemore
commonlyusedgroupfunctionsavailableinSQL.Notethatmostofthe
functionslistedcanperformaseitheranaggregateorananalyticfunction;afew
arelimitedtoonlyonetype.
TABLE7-1 CommonlyUsedAggregateFunctions—Overview
Likescalarfunctions,groupfunctionsacceptinput,processthatinput,and
returnoutput.Alsolikescalarfunctions,groupfunctionsacceptinputofspecific
datatypesandreturnoutputofspecificdatatypes,andtheinputandoutputdata
typesarenotnecessarilythesame,aswithscalarfunctions.Whilenumeric
groupfunctions(whichprocessandreturnnumericdata)arethemostcommon,
manygroupfunctionsprocessdatawithcharacteranddatedatatypes.
Groupfunctionsmustbetreatedseparatelyfromscalarfunctionsbecause
theybehavedifferently.AtypicalSELECTstatementcannotinvokegroup
functionsandscalarfunctionsatthesamelevelofaggregation.Forexample,
consideratableSHIP_SHOPwiththreerowsofdata.Ifweexecuteasimple
SELECTstatementonthetablewithascalarfunctionlikeROUND,wewillget
onerowreturnedforeveryrowprocessed.
IfweexecuteasimpleSELECTstatementagainstthesamesetofthreerows
usingagroupfunctionlikeAVG(foraverage),wewillgetaone-rowresult
showingtheaveragevalueforallthreerows.
Theproblemhereisthatwearestartingwiththreerowsofinputandtryingto
combineresultsofascalarandgroupfunctionatthesamelevelofaggregation.
Thescalar(ROUND)wantstoreturnthreerows,butthegroupfunction(AVG)
wantstoreturnonerow.ThisisnotpermittedtogetherinthesameSELECT
outputandtriggerstheerrormessage.
However,wecanusethetwotypesoffunctionsatdifferentlevelsof
aggregationinthesameSELECTstatement.Forexample,wecouldfirstperform
theaverage(thegroupfunction)andthenroundtheresults(thescalarfunction).
Alternatively,wecouldroundtheindividualvaluesfirstandthentakean
averageoftheroundednumbers.
Eitherway,thepointisthatwecannotapplybothscalarandgroupfunctions
togetherontorowsatthesamelevelofaggregation.However,wecanchooseto
nestthefunctioncallswithinasingleSQLstatementtoprocesseachsetofrows
oneatatime,passingtheresultsofonetoanother,combinedwithinasingle
SELECTstatement,bynestingthefunctions.
Let’slookatthedetailedfunctionalityofsomeofthemorecommonlyused
groupfunctions.
GroupfunctionscanbecalledfromfourplacesinaSELECTstatement:the
selectlist;theORDERBYclause;andeitheroftwoadditionalclauseswe’ll
lookatinthischapter,namely,theGROUPBYclauseandtheHAVINGclause.
Themajorgroupfunctionsaredescribedindetailinthefollowingsections.
Note:Allfunctionsarepresentedintheiraggregatefunctionsyntaxunless
otherwisespecifiedinthetext.
COUNT
Syntax:COUNT(e1)
Parameters:e1isanexpression.e1canbeanydatatype.
TheaggregatefunctionCOUNTdeterminesthenumberofoccurrencesof
non-NULLvalues.Itconsidersthevalueofanexpressionanddetermines
whetherthatvalueisNOTNULLforeachrowitencounters.
Forexample,considertheVENDORStable,showninFigure7-1.
FIGURE7-1 TheVENDORStable
Let’slookatalltherowsintheVENDORStable:
Here’stheoutput,showingthatwehavetworowsinthetable:
Wehaven’tselectedeverycolumnhere,andwedon’tneedtoforourpurposes.
ButassumethattheblankentriesintheoutputareNULLvaluesandnotblank
spaces—sothattheSTATUScolumncontainsonlyonevalue,andCATEGORY
hasnone.
Nowlet’slookatthefollowingSELECTstatementthattriestocount
occurrencesofdataineachofthesecolumns:
ThetermNULLdoesnotdescribeavalueatallbutratherthelackofavalue.
Itisthedatabasesaying,“Idon’tknowwhatthisvaluemightbe,ifanything.”
KeepthisinmindeachtimeyouencounterthephraseNULL:aNULListhe
lackofavalue.
NoticethatCOUNTignoresanyandallvaluesthatareNULL.Finally,we
cansimplycountthenumberofrowsintheentiretable.
COUNTwillreturnonlythenumber,orquantity,ofnon-NULLvaluesin
columns;or,whenusedtocountrows,itreturnsthenumberofrows.
RecallthatSELECT*FROMtableistheshorthandwayofaskingtoselect
allthecolumnsinagiventable.TheCOUNTfunctionisoftenusedwiththe
asteriskinthisfashiontogetaquickcountonalltherowsinagiventableusing
COUNT(*).
Wecouldhavemixedthesefunctionsinvariouscombinations.
NotethataCOUNToftheasteriskisaskingforacountofallrows.Inthe
raresituationwherearowcontainsnothingbutNULLvaluesinallofits
columns,COUNT(*)willstillcountthatrow.
It’sworthnotingthatCOUNTwillneverreturnNULL.Ifitencountersno
valuesatall,itwillatleastreturnavalueof0(zero).Thisisnotthecasewith
alloftheaggregates,butit’struewithCOUNT.Thisbecomesimportantwhen
workingwithsubqueries,whichwe’llstudyinChapter9.
TheDISTINCTandALLoperatorscanbeusedwithaggregatefunctions.
DISTINCTreturnsonlyuniquevalues.ALListheoppositeofDISTINCTandis
thedefaultvalue.IfyouomitDISTINCT,theALLisassumed.Inotherwords,if
youhaveneverusedDISTINCT,thenyou’vealwaysbeenimplyingALLupto
nowandthusqueryingalltherelevantresults.Soifyouhavethreerowsof
PRICEvaluesof2.99,3.99,and3.99,thenDISTINCTwouldreturnonly2.99
and3.99,whereALLwouldreturnallthreevalues,includingtheduplicates.
HereisanexampleshowingDISTINCTandALLusedwithinaCOUNT
function:
TheCOUNTfunctioncountsoccurrencesofdata,ignoringNULL
values.Butwhencombinedwiththeasterisk,asinSELECTCOUNT(*)
FROMVENDORS,itcountsoccurrencesofrows—andwillincluderows
withallNULLvaluesintheresults.
ThisexampletellsusthatthetablecalledEMPLOYEEShassevenrowswith
valuesforLAST_NAME,ofwhichfiveareuniquevaluesforLAST_NAME,so
twoareduplicates.AlsorememberthatDISTINCTandALLcannotbeused
withtheasterisk.
SUM
Syntax:SUM(e1)
Parameters:e1isanexpressionwhosedatatypeisnumeric.
TheSUMfunctionaddsnumericvaluesinagivencolumn.Ittakesonly
numericdataasinput.SUMaddsallthevaluesinalltherowsandreturnsa
singleanswer.Here’sanexample:
SuchaquerywilladdupallthevaluesforSUBTOTALintheORDERStable
andproduceasingleresult.Here’sanotherexample:
Thisquerywillfindanyandallrowsforwhichtheorderhasan
ORDER_DATEvaluesometimeinthemonthofApril2017andthenwilladdup
allthevaluesintheSUBTOTALcolumnandproduceasingleanswer.
MIN,MAX
Syntax:MIN(e1);MAX(e1)
Parameters:e1isanexpressionwithadatatypeofcharacter,date,ornumber.
ForagivensetofrowsidentifiedbyaSELECTstatement,MINreturnsthe
singleminimumvalue,andMAXreturnsthesinglemaximumvalue.MINand
MAXcanworkwithnumeric,date,andcharacterdata,andtheyusethesame
basiclogicthatORDERBYusesforthedifferentdatatypes,specifically:
NumericLownumbersareMIN;highnumbersareMAX.
DateEarlierdatesareMIN;laterdatesareMAX.Earlierdatesare
lessthanlaterdates.
CharacterAislessthanZ;Zislessthana.Thestringvalue2is
greaterthanthestringvalue100.Thecharacter1islessthanthe
characters10.
NULLvaluesareignored,unlessallvaluesareNULL,inwhichcaseMINor
MAXwillreturnNULL.
Forexample,considerthefollowinglistofdatafromthetableEMPLOYEES:
Nowlet’sidentifyMINandMAXvalues.
NotethatHoddlesteinisalphabeticallythefirstvaluefromthelistof
LAST_NAMEvalues.
EventhoughthedatareturnedbyMINandMAXrepresentsthedatafound
withinasinglerowinthelist,donotbetrickedintothinkingthatthisrepresents
asingle-rowanswer—itdoesnot.SQLseeseachresponseofMINandMAXas
anaggregateanswer,meaningthattheindividualvalueistheanswer
representingthefullsetofrows.
AVG
Syntax:AVG(e1)
Parameters:e1isanexpressionwithanumericdatatype.
TheAVGfunctioncomputestheaveragevalueforasetofrows.AVGworks
onlywithnumericdata.ItignoresNULLvalues.Forexample,let’slookatthe
PAY_HISTORYtable;afterthat,we’llaskfortheaveragevalueofallthevalues
withintheSALARYcolumn.
Whilewe’reatit,wecannesttheresultsofthisquerywithinthescalar
functionROUND,likeso:
WecangetreallyfancyandformatthedatausingtheTO_CHARfunction
andaformatmodel.
IntheselastfewexamplesofSELECTstatements,we’venestedasingle
aggregatefunctionwithintwoscalar,orsingle-row,functions.Youcan
incorporateasingleaggregatefunctionwithinasmanynestedscalarfunctionsas
youwant.Theaggregatefunctionneednotbetheinnermostfunction;youcan
includeoneaggregatefunctionwithanynumberofscalarfunctionsinanested
combination,providedthatalloftheparameterdatatypesarerespected.Butif
youwanttoincludetwoaggregatefunctionswithinanestedcombination,be
careful—therearelimitationsontheuseofnestedaggregatefunctions.Wewill
addressthatissuelaterinthischapter.It’smorecomplexthanitmightappear.
DISTINCTandALLareavailableforusewithAVG.Intheeventthata
table’sdatalistingincludessomerepeatedvalues,theuseofDISTINCTwill
transformtheresultssothattheaverageiscomputedonlyonuniqueoccurrences
ofeachvalue.
MEDIAN
Syntax:MEDIAN(e1)
Parameters:e1isanexpressionwithanumericordatedatatype.
MEDIANcanoperateonnumericordatedatatypes.ItignoresNULLvalues.
TheMEDIANfunctionissomewhatrelatedtoAVG.MEDIANperformsasyou
mightexpect:fromasetofdata,MEDIANreturnseitherthemiddlevalueor,if
thatisn’teasilyidentified,thenaninterpolatedvaluefromwithinthemiddle.In
otherwords,MEDIANwillsortthevalues,andifthereisanoddnumberof
values,itwillidentifythevalueinthemiddleofthelist;otherwise,iftherean
evennumberofvalues,itwilllocatethetwovaluesinthemiddleofthelistand
performlinearinterpolationbetweenthemtolocatearesult.
Here’sanexample:
IfyouweretoexecutethepreviousSQLstatements,thevaluereturnedbythe
SELECTstatementwouldbe3.
RANK
RANKhastodowiththenumericrank(1,2,3,andsoon)ofavaluewithina
largegroupofvalues.Whenatiedsetofvaluesamongrowsisencountered,
RANKassignsthesameranknumbertothetiedrowsbutkeepscountofeach
rowandresumescountingsubsequentrowswiththecorrectlynumbered
position.
TheanalyticversionofRANKreturnstherankingsofasetofrows.The
aggregateversionhasoneparameterofanexpressionpertainingtoarowand
returnstherankforthatvalue.BothversionsofRANKaredescribedinthe
followingsections.
RANK:Analytic
Syntax:RANK()OVER(PARTITIONBYp1ORDERBYob1)
Parameters:p1isapartition.ob1isanexpression.
TheuseofPARTITIONBYisoptional.Allotherelementsarerequired.
TheRANKfunctioncalculatestherankofavaluewithinagroupofvalues.
RanksmaynotbeconsecutivenumberssinceSQLcountstiedrowsindividually,
soifthreerowsaretiedforfirst,theywilleachberanked1,1,and1,andthe
nextrowwillberanked4.
Forexample,notetheuseofRANK()inthefollowingline2(linenumbers
added):
Intheexample,thevaluesreturnedbyRANKareinthecolumnSQ_FT_RK.
TherowsarepartitionedbyROOM_STYLE,sothosearesortedfirst;then
withinagivenpartition,rowsarerankedbySQ_FT.Noteline14,whichisthe
startofanewROOM_STYLE,andtheRANKvaluerestartsat1.Alsonotethat
thenextline,line15,hasthesamevalueforSQ_FT,sotherankingistiedat1.
Next,line16isranked3sincethisisthethirdrowof'Suite'ROOM_STYLES,
andtheSQ_FTof586isnottiedwithanyotherrow.
RANK:Aggregate
Syntax:RANK(c1)WITHINGROUP(ORDERBYe1)
Parameters:c1ande1areexpressions.
Inthisformat,theparameterscanberepeatedinsuchawaythatforeachc1,
youcanhaveacorrespondinge1;foreachc2(ifincluded),theremustbea
correspondinge2;andsoon.Eachsuccessiveparameterisseparatedfromthe
previousparameterbyacomma,asin
Also,thedatatypeofc1mustmatchthedatatypeofe1,thedatatypeofc2
(ifincluded)mustmatchthedatatypeofe2,andsoon.
Here’sanexample:
TheexampleshowsthatforaSQ_FTvalueof533,therankingis4.
DENSE_RANK
DENSE_RANKissimilartoRANK.DENSE_RANKreturnsthenumericrank
(1,2,3,andsoon)ofavaluewithinalargegroupofvalues.Whenatieis
encountered,DENSE_RANKassignsthesamenumbertoeachequivalentvalue.
Butwhenitresumescountingafteratie,thisiswhereDENSE_RANKdiffers
fromRANK.TheDENSE_RANKfunctionwillnotskipanynumbers,andit
willassignthenextsequentialnumber,regardlessofhowmanytied(equivalent)
valuesitjustnumbered.
DENSE_RANK:Analytic
Syntax:DENSE_RANK()OVER(PARTITIONBYp1ORDERBYob1)
Parameters:p1isapartition.ob1isanexpression.
InthecaseofDENSE_RANK,ifthreerowsaretiedforfirst,theywilleach
beranked1,1,and1,andthenextrowwillberanked2.
Forexample,notetheuseofDENSE_RANK()inthefollowingline2:
Notelines16and17.WhenweusedRANK,theranksherewere3and4,
respectively.UsingDENSE_RANK,therankingsare2and3.Oncethetwotied
rowsonlines14and15arereturned,DENSE_RANKresumesranking
subsequentrowswiththenextsequentialnumber,regardlessofhowmanytied
rowsmayhavepreceded.
DENSE_RANK:Aggregate
Syntax:DENSE_RANK(c1)WITHINGROUP(ORDERBYe1)
Parameters:c1isaconstant;e1isanexpressionwithadatatypematchingthe
correspondingc1datatype.Numericandcharacterpairsareallowed.
TherulesfortheaggregateformofDENSE_RANKarethesameasRANK’s
aggregateform.ThedifferenceisthatDENSE_RANKwillreturnavalue
consistentwiththeDENSE_RANKlogicasexplainedintheprevioussection;
thatis,ifaseriesoftiedrowsisencountered,ranknumbersofsubsequentrows
willbeassignedwiththenextsequentialnumber,regardlessofhowmanytied
rowspreceded.So,iffourrowsaretiedforsecond,theywilleachberanked2,
2,2,and2,andthenextrowwillberanked3.
FIRST,LAST
Syntax:aggregate_functionKEEP(DENSE_RANKFIRSTORDERBYe1)
aggregate_functionKEEP(DENSE_RANKLASTORDERBYe1)
Parameters:e1isanexpressionwithanumericorcharacterdatatype.
ThefunctionsFIRSTandLASTaresimilar.Bothareconsideredaggregate
functionsaswellasanalyticfunctions.Foragivenrangeofsortedvalues,each
returnseitherthefirstvalue(FIRST)orthelastvalue(LAST)ofthepopulation
ofrowsdefininge1,inthesortedorder.
Here’sanexample:
Inthisexample,wearedoingthefollowing:
First,we’resortingalltherowsintheSHIP_CABINStable
accordingtothevalueintheGUESTScolumn,andwe’reidentifyingthe
FIRSTvalueinthatsortorder,whichisacomplexwayofsayingthat
we’reidentifyingthelowestvaluefortheGUESTScolumn.
ForallrowswithaGUESTvaluethatmatchesthelowestvaluewe
justfound,determinetheMAXvalueforSQ_FT.
Inothers,displaythehighestnumberofsquarefeetforanyandall
cabinsthataccommodatethelowestnumberofguestsaccordingtothe
GUESTScolumn.
ExperiencedprofessionalsmightrecognizethatFIRSTandLASTperform
tasksthatcanalsobedonewithcertainusagesofself-joinsorviews,whichwe
examineinlaterchapters.Whileself-joinsandviewsarebeneficialfora
varietyofreasons,theuseofFIRSTorLASTasshownpreviouslywillachieve
performanceimprovementsoverthealternativeapproaches.
Others
Therearemoreaggregatefunctionsthanaredescribedinthissection.They
includefunctionstoworkwithnestedtables,functionstoperformlinear
regressionanalysis,andvariousformsofstatisticalanalysis.Thesefunctions
aren’tspecificallyreferencedinthecertificationexamguideobjectives,sowe
won’treviewthemallhere.ButyoucanfindfulldescriptionsoftheminOracle
Corporation’sSQLLanguageReferenceManual.
CERTIFICATIONOBJECTIVE7.02
GroupDatabyUsingtheGROUPBYClause
TheGROUPBYclauseisanoptionalclausewithintheSELECTstatement.Its
purposeistogroupsetsofrowsandtreateachindividualsetasawhole.Inother
words,GROUPBYidentifiessubsetsofrowswithinthelargersetofrowsbeing
consideredbytheSELECTstatement.Inthisway,it’ssortoflikecreatinga
seriesofmini-SELECTstatementswithinthelargerSELECTstatement.
GROUPBYisuniquetotheSELECTstatement;itdoesnotexistinother
SQLstatements.
Let’stakeanotherlookattheSHIP_CABINStable,whichnowhassomenew
columnssincethelasttimeweworkedwithit(seeFigure7-2).
FIGURE7-2 TheSHIP_CABINStable
Let’srunthisSELECTstatementagainstthetable,lookingonlyatrows
whereSHIP_ID=1:
Theresultsareshownhere:
TakealookatthecolumncalledSQ_FT,showingthenumberofsquarefeet
ofeachroomontheship.Let’scomputetheaveragesquarefeetforroomsand
roundofftheanswer:
That’stheaverageforallofthecabinsontheship.Butlookatthedata
listing,andyou’llseethateachoftheship’scabinsseemstofallintooneoftwo
differentcategoriesaccordingtothedataintheROOM_STYLEcolumn.Each
roomiseithera'Suite'ora'Stateroom'.
Ifwewantedtolookattheaverageforjustthesuitesandalsoforjustthe
staterooms,wecouldruntwoindividualqueries,likethis:
Thatisusefulinformation,butit’sarelativelycumbersomewaytogetit.
Usingthisapproach,wehaveto(a)identifytheindividualvaluesfor
ROOM_STYLEandtypethemcarefullyintoourqueries,(b)runmultiple
queries,and(c)obtainouroutputviamultiplequeries.
ThebetterwaytogetthisdoneiswiththeGROUPBYclause.Wecangetthe
sameinformationbytellingSQLtogrouptherowsaccordingtotheirvaluesfor
ROOM_STYLE,whatevertheymaybe.Here’sthequery:
Inthisparticularexample,weaddtheGROUPBYclausetotellSQLtogroup
therowsthathavethesamevalueforROOM_STYLE,andthenwecomputethe
AVGfunctionforeachgroup,ratherthanforalltherowsinthetable.Notethat
we’restillusingtheWHEREclause,soweonlyaddressrowswithaSHIP_ID
valueof1.
TogetanideaofwhatSQLdoeswiththisquery,let’sfirstsorttherows
accordingtoROOM_STYLEandhighlightthetwodifferentgroupsofrows.
OurGROUPBYquerydidn’tincludeanORDERBYclause,butwechoseto
sorttheserowstohighlightthefactthattherearetwogroupsofrows.
NowgobackandlookatourSELECTstatement.Didwespecifyanything
about'Suite'or'Stateroom'?Notspecifically.Wedidn’thaveto.Thedirectiveto
GROUPBYROOM_STYLEtellsSQLtogroupeachsetofrowsthatsharethe
samevalueforROOM_STYLE,whateverthatmaybe.
Wecouldhaveincludedmultipleaggregatefunctionsinthisquery’sselectlist
ifwewanted.Here’sanexample:
TherulesforformingaGROUPBYclauseareasfollows:
TheGROUPBYcanspecifyanynumberofvalidexpressions,
includingcolumnsofthetable.
GenerallytheGROUPBYisusedtospecifycolumnsinthetable
thatwillcontaincommondatainordertogrouprowstogetherfor
performingsomesortofaggregatefunctiononthesetofrows.
Thefollowingaretheonlyitemsallowedintheselectlistofa
SELECTthatincludesaGROUPBYclause:
ExpressionsthatarespecifiedintheGROUPBY.
Aggregatefunctionsappliedtothoseexpressions.
ExpressionsthatarespecifiedintheGROUPBYdonothavetobe
includedintheSELECTstatement’sselectlist.
Let’strygroupingthissamesetofrowsbysomethingelse.Inthisquery,
we’llgroupbytheROOM_TYPEcolumninstead.We’lladdafewotherfeatures
aswell.
Noticethefollowingchangestoourquery:
Aswestated,wechosetogroupbytheROOM_TYPEcolumn.We
alsoputROOM_TYPEintheselectlist.
WeaddedanORDERBYclausethatissortingonthesecond
columnfromtheselectlist,whichinthiscaseistheAVGoftheSQ_FT
column.
WereplacedtheMINfunctionwiththeMAXfunction,justforfun.
UnrelatedtotheGROUPBYfunctionality,wechosetoapplya
formatmodeltotheAVGoutputtocleanitupalittle.
Wealsoaddedacolumnaliasforeachofthelastthreeexpressions,
omittingtheoptionalkeywordASforeachcolumnalias.
NoticetheresultsofourmodifiedSELECTwithGROUPBYclause:
ThevaluesforROOM_TYPEareautomaticallylisted,andinthis
case,fivevalueswerefound—sowehavefiverowsinouroutput,each
representingasetofrowsinthesourcetable.
TheaggregatefunctionsofAVG,MAX,andCOUNTareall
calculatedforeachindividualgroup.
Thatlastpointisimportant.It’stheentirepurposeoftheGROUPBY
function.Ifyoudon’tunderstandit,thentrythis—lookattheoutputofthefirst
row,whichisfortheROOM_TYPEvalueof'Royal',andconsiderthatthe
individualrowyouareseeingintheoutputisthesamedatayouwouldgetifyou
ranthisqueryalone,withouttheGROUPBYclause:
Intheprecedingquery,we’veeliminatedtheGROUPBYclauseand
introducedaWHEREclausetolookonlyatROOM_TYPE='Royal'.Theresult
isthesamedatawefindinthefirstrowoftheGROUPBYweranearlier,except
forthevalueoftheROOM_TYPEcolumn,whichwas'Royal'.Ifyourepeated
thisqueryfourmoretimesandchangedtheROOM_TYPEvalueeachtimeto
theotherfourroomtypevalues('Presidential','Skyloft','Standard',and'Large'),
youwouldeventuallyre-createthesameresultsoftheonesingleGROUPBY
statementweexecutedearlier.
Groupingbyaparticularcolumntransformsthatcolumnintoanaggregate
value—ifonlytemporarilyforthepurposeofthequery.
MultipleColumns
YoucangroupmultiplecolumnswithasingleGROUPBYclause.
Herearesomedetailsoftheprecedingexample:
TheGROUPBYclauseincludestwocolumnsfromthetable:
ROOM_STYLEandROOM_TYPE.Theorderisimportant:ittellsSQL
togroupallrowsfirstthatsharethesamevaluefortheROOM_STYLE
and,foreachsetofrowswiththesameROOM_STYLE,grouprows
withthesameROOM_TYPEvalue.Thereareapparentlysixsuch
groups,accordingtotheoutput.
TheSELECTstatement’sselectlisthappenstoincludethe
ROOM_STYLEandROOM_TYPEcolumnsinthesamepositionsas
theyareintheGROUPBYclause;thisisnotrequired,butitmakesthe
outputlistingeasytoread.
TheORDERBYclausetellsSQLtosorttherowsbasedonthevalue
inthethirditemintheselectlist,whichistheMINaggregatefunction.
EachoftheaggregatefunctionsisformattedwiththeTO_CHAR
formatmodeltoincludecommaswhereappropriateandnarrowthe
columnstoareasonablewidth.
Thefinalcolumninourselectlistisanexpressionthatcalculatesthe
MINandMAXvalues,determinesthedifferencebetweenthem,and
formatstheresults.
Clearlythere’salotgoingoninthisSELECTstatement,butit’sagreat
exampleofaGROUPBYinaction.
ORDERBYRevisited
WhenaGROUPBYisusedinaSELECTstatement,thenifthereisanORDER
BYclauseincludedaswell,itsusewillbesomewhatrestricted.Thelistof
columnsorexpressionsinanORDERBYthatispartofaSELECTstatement
thatusesGROUPBYislimitedtothefollowing:
ExpressionsspecifiedintheGROUPBYclause
Expressionsspecifiedintheselectlist,referencedbyposition,name,
oralias
Aggregatefunctions,regardlessofwhethertheaggregatefunctionis
specifiedelsewhereintheSELECTstatement
ThefunctionsUSER,SYSDATE,andUID
AsingleSELECTstatementmustproduceoutputatjustonelevelof
aggregation.ThisiswhyaSELECTstatementcannotmixscalarand
aggregatevaluesinaselectlist,unlessitistonestthemwithineach
other.
NotethatwhenaGROUPBYisused,theORDERBYclausemustreference
eithercolumnsinthetablethatarespecifiedintheGROUPBYclauseorthe
outputofanaggregatefunctionorboth.Otherreferencesarenotpermitted.
That’snotthecaseforSELECTstatementsingeneral.InascalarSELECTyou
canuseORDERBYoncolumnsinthetablewhethertheyareincludedinthe
SELECTornot.Butthat’snottruewhenaGROUPBYisinvolved.ORDERBY
ismorelimited.
NestingFunctions
Youmightrecallfromourexamplesthatwe’venestedfunctionsinsomeofour
SQLstatements.Theconceptofnestingfunctionsreferstothepracticeof
positioningafunctioninsuchawaythatthevalueitreturnsbecomestheinput
parameterforanotherfunction.Here’sanexample:
Inthisexample,theMEDIANaggregatefunctionisnestedwithinthe
TO_CHARfunction.Thus,theMEDIANfunction,inthisexample,istheinner
function,anditreturnsavaluethatbecomesaninputparametertothe
TO_CHARconversionfunction,whichistheouterfunction.Aslongasthedata
typesmatchup,nestingisallowed.Inthisinstance,MEDIANcanbeusedlike
thisonlyifitisreturningavalueofthedatatypethatTO_CHARcanaccept.
Now,rememberthattherearetwogeneraltypesoffunctions:single-row,or
scalar,functions;andmultirow,oraggregate,functions.Scalarfunctionsreturn
onevalueforeachrowencounteredbytheSQLstatementinwhichthescalar
functionisapplied.Aggregatefunctionsreturnonevalueforeveryzeroormore
rowsencounteredbytheSQLstatement.
Therulesfornestingfunctionsdiffer,dependingonwhetheryouarenesting
aggregateorscalarfunctions.
Scalarfunctionscanbenestedmultipletimes,aslongasthedatatypesmatch.
Thereasonisthatscalarfunctionsalloperateonthesamelevelofdetailwithin
therows—foreveryonerow,ascalarfunctionreturnsonevalue.
Aggregatefunctions,ontheotherhand,behavedifferentlythanscalar
functionswhenitcomestonesting.Thereasonisthataggregatescombinedata
frommultiplerowsintoasinglerow.Oncethathasbeenaccomplished,your
resultingvaluenolongerrepresentsthesamelevelofdetailyouwereoriginally
dealingwith.
Forexample,let’slookagainattheSHIP_CABINStableandconsidera
simpleSELECTstatementwithaGROUPBYclause:
ThisSELECTgroupsrowsaccordingtoROOM_STYLEandROOM_TYPE,as
we’veseen,anddisplaystheMAXvalueforSQ_FTwithineachgroup.The
resultissixrows,whichtellsusthatwehavesixuniquegroupsfor
ROOM_STYLEandROOM_TYPE.Ittellsusnothingaboutthenumberof
rowsthatmightbefoundwithinthesourcedata—butjustsoyouknow,it’sthe
samedatalistwesawearlierinthischapter,whichconsistedof12rows.
Nowlet’strytocomputetheAVGvalueofthesesixMAXvalues:
IfwetrytoexecutethisSELECTstatement,thiswillbetheresult:
Whyarewegettingthiserror?Thereasonissimple:byintroducingAVGinto
theSELECTstatement’sexpressionlist,wearemovingupthelevelof
aggregationahigherdegree,andweareinformingSQLthatwesimplywantone
answerforallofthosegroupedrows.Theproblem:ourGROUPBYistryingto
displayonlyoneanswerforalltherows,buttherearesixdifferent
ROOM_STYLEandROOM_TYPEvaluesforthoserows.Inotherwords,there
isnosingleanswerthatcanrepresentallsixofthoserows.
OnesolutionistomodifyourSELECTstatementbyremovingitemsfromthe
selectlistthatareatinconsistentlevelsofdetail,likethis:
Nowwegetananswer,andwecanseetheresultdisplaysinjustonerow,
whichinturnrepresentsthesixrowsfromthepreviousquery,whichitself
represented12rows.So,wehaveaggregated12rowsinto6andthen6into1—
allwithonequery.Inotherwords,thislatestSELECTstatementanditsAVG
resultrepresentanaggregationofanaggregation.
Let’stryathirdlevelofnestedaggregate:
Itcan’tbedone.Twolevelsdeepisthefurthestyoucangowithnested
aggregatefunctions.
However,weareallowedtointroducenestedscalarfunctionsatanytime.
Here’sanexample:
Tosumup:
Youareallowedtonestaggregatefunctionsuptotwolevelsdeep.
Eachtimeyouintroduceanaggregatefunction,youarerollingup
lower-leveldataintohigher-levelsummarydata.
YourSELECTstatement’sselectlistmustalwaysrespectthelevelof
aggregationandcanonlyincludeexpressionsthatareallatthesame
levelofaggregation.
Andfinally,rememberthatscalarfunctionscanbenestedatanytimeand
havenoeffectonmodifyingthelevelsofrowaggregation.
CERTIFICATIONOBJECTIVE7.03
IncludeorExcludeGroupedRowsbyUsingthe
HAVINGClause
TheHAVINGclausecanexcludespecificgroupsofrowsdefinedintheGROUP
BYclause.YoucouldthinkofitastheWHEREclauseforgroupsofrows
specifiedbytheGROUPBYclause.
TheHAVINGclauseisuniquetotheSELECTstatement;itdoesnotexistin
otherSQLstatements.
TheHAVINGclausedoesnotdefinethegroupsofrowsthemselves;those
groupsmustalreadybedefinedbytheGROUPBYclause.HAVINGdefinesthe
criteriauponwhicheachoftheGROUPBYgroupswillbeeitherincludedor
excluded.
TheHAVINGclausecanbeinvokedonlyinaSELECTstatementwherethe
GROUPBYclauseispresent.
Ifitisincluded,GROUPBYandHAVINGmustfollowWHERE(if
included)andprecedeORDERBY(ifincluded).Table7-2showsthese
relationships.
TABLE7-2 ClausesinaSELECTStatement
Here’sanexample:
NotetheHAVINGclauseonlines7and8.ThisparticularHAVINGrestricts
thegroupsidentifiedintheGROUPBYclausetoonlythoserowswhere
ROOM_TYPE='Standard'orROOM_TYPE='Large'orthevaluefor
MIN(SQ_FT)isgreaterthan1200.
Astheprecedingexampleshows,HAVINGcanusethesameBoolean
operatorsthatWHEREdoes—AND,OR,andNOT.Thesearetheonly
restrictionsonHAVING:
ItcanbeusedonlyinSELECTstatementsthathaveaGROUPBY
clause.
Itcanonlycompareexpressionsthatreferencegroupsasdefinedin
theGROUPBYclauseandaggregatefunctions.
HAVINGcanincludescalarfunctionsaslongastheserestrictionsare
respected.Inotherwords,ascalarfunctioncanbeincorporatedintoalarger
expressionthatreferencesaGROUPBYgrouporanaggregatefunction.Just
rememberthatHAVINGdealswithgroupsofrows,notindividualrows.
CERTIFICATIONSUMMARY
Multirowfunctionsdifferfromsingle-rowfunctionsinthattheyreturnonevalue
foragroupofrows,whereassingle-rowfunctionsreturnoneanswerforeach
individualrow.Single-rowfunctionsarealsoreferredtoasscalarfunctions;
multirowfunctionsarealsoreferredtoasaggregatefunctions.Theavailable
aggregatefunctionsinSQLincludethecommonlyusedfunctionsCOUNT,
SUM,MIN,MAX,AVG,andMEDIAN.Thereareaggregatefunctionsto
supportstandarddeviationandvariance,ranking,linearregressionanalysis,
groupingwithROLLUPandCUBE,XMLsupport,andworkingwithnested
tables.
AggregatefunctionscanbecalledfromfourplacesinaSELECTstatement:
theselectlist,theGROUPBYclause,theHAVINGclause,andtheORDERBY
clause.
IfanaggregatefunctionappearsintheselectlistoftheSELECTstatement,
thenallotherexpressionsmustbeatthesamelevelofaggregation.Youcannot
use,inasingleSELECTstatement,ascalarfunctionononecolumnandan
aggregatefunctiononanother.TheSELECTmustbeabletoreturnrowdatathat
iscollectivelyatthesamelevelofdetail.
TheGROUPBYclausespecifiesoneormoreexpressionsthatSQListouse
togrouprowstogether.AnyvaluesdisplayedintheSELECToutputmustbe
displayedonceforeachgroup.Inthisfashion,theGROUPBYcantransforma
singlecolumnreferenceintoanaggregatereferencebyspecifyingthatcolumnas
aGROUPBYitem.
AnycolumnorexpressionspecifiedintheGROUPBYclausecanalsobe
includedintheSELECTstatement’sselectlist.However,thisisnotrequired.
TheHAVINGclausecanbeusedtofilteroutgroupsthathavebeenspecified
withtheGROUPBYclause.HAVINGworksmuchliketheWHEREclause:
whiletheWHEREclausefiltersoutindividualrows,HAVINGdoesthesame
thingforgroupsofrows.HAVINGisallowedonlyiftheGROUPBYclauseis
present.
IfHAVINGandGROUPBYareincludedinaSELECTstatement,theymust
followtheWHEREclause(ifused)andprecedetheORDERBYclause(if
used).GROUPBYoftenprecedesHAVING,butHAVINGispermittedto
precedeGROUPBYintheSELECTstatement’ssyntax.
✓TWO-MINUTEDRILL
DescribetheUseofGroupFunctions
Groupfunctionsarealsoknownasaggregate,ormultirow,
functions.
Multirowfunctionsreturnonevalueforeverysetofzeroormore
rowsconsideredwithinaSELECTstatement.
Therearegroupfunctionstodetermineminimumandmaximum
values,calculateaverages,andmore.
Groupfunctionscanbeusedtodeterminerankwithinagroupof
rows.
AggregateandscalardatacannotbeincludedinthesameSELECT
statement’sselectlist.
TheCOUNTfunctioncountsoccurrencesofdata,asopposedtothe
SUMfunction,whichaddsupnumericvalues.
TheMINandMAXfunctionscanoperateondate,character,or
numericdata.
TheAVGandMEDIANfunctionscanperformaverageandmedian
calculations,andtheycanignoreNULLvaluesintheircomputations.
SomefunctionssuchasRANKusethekeywordsWITHINGROUP
(ORDERBY)toprocessavalueandidentifyitsrankingwithinthe
overallsetofrowsinthedataset.
GroupDatabyUsingtheGROUPBYClause
TheGROUPBYclauseisanoptionalclauseintheSELECT
statementinwhichyoucanspecifyhowtogrouprowstogetherinorder
toprocessthemasagroup.
TherowgroupsidentifiedbyGROUPBYcanhaveaggregate
functionsappliedtothemsothatthefinalresultoftheSELECTisnota
singleaggregatevaluebutaseriesofaggregatefunctionresults,oneper
group.
GROUPBYcanspecifycolumnsinatable,whichwillhavethe
effectofgroupingrowsintheSELECTthatsharethesamevaluesin
thosecolumns.
WhateveryouspecifyintheGROUPBYmayalsobeincludedin
theSELECTstatement’sselectlist—butthisisnotrequired.
TheeffectofGROUPBYonacolumnistochangethatcolumn
intoanaggregatevalue;inotherwords,bygroupingrowsthathave
commondataforagivencolumnandbyspecifyingthecolumninthe
GROUPBY,youelevatetheinformationinthecolumnfromscalarto
aggregateforthepurposesofthatSELECTstatement.
GROUPBYcanspecifyoneormoreexpressions.
IncludeorExcludeGroupedRowsbyUsingtheHAVINGClause
TheHAVINGclauseisanoptionalclausefortheSELECT
statementthatworksincoordinationwithGROUPBY.
YoucannotuseHAVINGunlessGROUPBYispresent.
HAVINGspecifiesgroupsofrowsthatwillbeexcludedinthe
outputoftheSELECTstatement.
HAVINGperformsthesamefunctionforGROUPBYthatWHERE
performsfortherestoftheSELECTstatement.
HAVINGspecifiesgroupsusingthesameexpressionlogicand
syntaxthatWHEREwoulduse.
TheGROUPBYandHAVINGclausesarenotrequired,butifused,
theymustfollowtheWHEREclauseandprecedetheORDERBY
clause.
WhileGROUPBYtypicallyprecedesHAVINGincommon
practice,thisisnotrequired,andtheycanappearineitherorder;
HAVINGcanprecedeGROUPBY.
SELFTEST
Thefollowingquestionswillhelpyoumeasureyourunderstandingofthe
materialpresentedinthischapter.Chooseonecorrectanswerforeachquestion
unlessotherwisedirected.
DescribetheUseofGroupFunctions
1.Whichofthefollowingistrueaboutaggregatefunctions?(Choose
two.)
A.ReturnonevalueforeachgroupofrowsspecifiedinaSELECT
statement.
B.Arealsocalledgroupfunctions.
C.Willcausearun-timeerrorwhenusedinSELECTstatements
thatreturnzerorowsoronerow.
D.Canoperateonlywithnumericdata.
2.Reviewthefollowingillustration:
NowreviewthisSQLstatement:
Whatcanbesaidofthisstatement?
A.ItwillfailtoexecutebecauseORDER_DATEisadatedatatype,
andnoaggregatefunctioncanworkwithadatedatatype.
B.Itwillfailtoexecutebecauseitmixesscalarandaggregatedata
intheselectlist.
C.Itwillexecutesuccessfullybutnotproduceanymeaningful
output.
D.ThereisnothingwrongwiththeSQLstatement.
3.Whichofthefollowingaggregatefunctionscanbeusedoncharacter
data?(Choosetwo.)
A.COUNT
B.MIN
C.AVG
D.MEDIAN
4.ExaminethefollowingdatalistingofatablecalledPERMITS:
Whichoneofthefollowingaggregatefunctionscouldbeusedto
determinehowmanypermitshavebeenfiledbyVENDOR_ID101?
A.SUM
B.COUNT
C.MEDIAN
D.HAVING
5.Reviewtheillustrationfromquestion2,andthenreviewthefollowing
SQLstatement:
WhatwillresultfromanattempttoexecutethisSQLstatementonthe
CRUISE_ORDERStable?
A.Itwillfailwithanexecutionerrorbecauseyoucannotusethe
AVGfunctiononaPRIMARYKEYcolumn.
B.Itwillfailwithanexecutionerrorbecauseyoucannotusethe
MINfunctiononaDATEdatatype.
C.Itwillfailwithanexecutionerrorifthetablecontainsonlyone
row.
D.Itwillexecuteandperformasintended.
6.ReviewthefollowingdatalistingfromatableSCORES:
Nowconsiderthefollowingquery:
Whatwillbetheresultofthisquery?
A.ItwillresultinasyntaxerrorbecauseoftheTO_CHAR
function.
B.Itwillresultinanexecutionerror.
C.90.00.
D.60.00.
7.Reviewthefollowingillustration:
WhichofthefollowingSQLstatementswillexecutecorrectly?
A.SELECTRANK(100000)WITHINGROUP(ORDERBY
PROJECT_COST)FROMPROJECTS;
B.SELECTRANK(100,000)WITHINGROUP(ORDERBY
PROJECT_COST)FROMPROJECTS;
C.SELECTRANK(7500000)GROUPBY(ORDERBY
PROJECT_COST)FROMPROJECTS;
D.SELECTRANK('Upgrade')WITHINGROUP(ORDERBY
PROJECT_COST)FROMPROJECTS;
8.WhichofthefollowingaggregatefunctionsignoresNULLvaluesin
itscalculations?(Chooseallthatapply.)
A.MEDIAN
B.AVG
C.SUM
D.MAX
9.Anaggregatefunctioncanbecalledfromwithin:(Choosetwo.)
A.TheHAVINGclauseofanINSERTstatement
B.TheORDERBYclauseofaSELECTstatement
C.TheexpressionlistofaDELETEstatement
D.TheselectlistofaSELECTstatement
GroupDatabyUsingtheGROUPBYClause
10.Reviewtheillustrationfromquestion7.YourtaskistodefineaSELECT
statementthatgroupsrowsaccordingtotheirvalueforPURPOSEand,for
eachpurpose,addsupthevaluesstoredinDAYS.Whichoneofthe
followingquerieswillperformthistask?
11.Reviewtheillustrationfromquestion2,andthenlookattheSQLcodethat
follows:
Recallthatthe'Q'formatmodelisforquarter,soTO_CHARusinga
DATEdatatypewiththe'Q'formatmaskistranslatingthedateintothe
quarterinwhichitfalls—1,2,3,or4.Giventhat,whichofthefollowing
statementsistrueoftheSQLstatement?
A.Itwillfailbecauseofasyntaxerrorinline4sinceyoucannot
usetheTO_CHARfunctionintheGROUPBYclause.
B.Itwillfailbecauseofasyntaxerrorinline1sinceyoucannot
usetheTO_CHARfunctionwiththeCOUNTaggregatefunction.
C.Itwillexecuteandshowthenumberofordersinthe
CRUISE_ORDERStableforeachquarterintheyear2009.
D.Noneoftheabove.
12.Reviewtheillustrationfromquestion7,andthenlookattheSQLcodethat
follows:
WhatwillhappenifyoutrytoexecutethisqueryonthePROJECTS
table?
A.Itwillfailwithasyntaxerrorbecauseline1isnotcorrect.
B.Itwillfailwithanexecutionerrorbecauseyoucannotusea
VARCHAR2columninaGROUPBYclause.
C.Itwillsucceedanddisplayonerowforeachdifferentvaluein
thePURPOSEcolumn.
D.Itwillsucceedanddisplayonerow.
IncludeorExcludeGroupedRowsbyUsingtheHAVINGClause
13.WhichofthefollowingstatementsistrueaboutHAVING?(Choosetwo.)
A.ItcanbeusedonlyintheSELECTstatement.
B.ItmustoccuraftertheGROUPBYclause.
C.ItmustoccuraftertheWHEREclause.
D.Itcannotreferenceanexpressionunlessthatexpressionisfirst
referencedintheGROUPBYclause.
14.Reviewtheillustrationfromquestion7,andreviewtheSQLstatementthat
follows:
WhichofthefollowingstatementsistrueforthisSQLstatement?
A.Itwillfailtoexecutebecauseofasyntaxerroronline4.
B.ItwillincludeonlythoserowswithaPROJECT_COSTvalueof
lessthan500000.
C.ItwillincludeonlythosegroupsofrowsforagivenSHIP_ID
withanaveragevalueofPROJECT_COSTlessthan500000.
D.Itwillfailtoexecutebecauseofasyntaxerroronline1.
15.Reviewtheillustrationfromquestion7.Yourassignment:createaSELECT
statementthatqueriesthePROJECTStabletoshowtheaverageproject
costforeachPURPOSE.Youknowthereareonlytwovaluesfor
PURPOSEinthetable:'Upgrade'and'Maintenance'.Youwanttorestrict
outputtothoserowswhereDAYSisgreaterthan3.Whichofthefollowing
SELECTstatementswillperformthistask?
SELFTESTANSWERS
DescribetheUseofGroupFunctions
1. AandB.Aggregatefunctionsreturnonevalueforeachgroupof
rowsintheSELECTstatement.Theyarealsoreferredtoasgroup
functions,ormultirowfunctions.
CandDareincorrect.AGROUPBYcanbeusedinaSELECT
statementthatreturnszerorows,onerow,ormultiplerows.Someaggregate
functionsoperateoncharacteranddatedatatypes;theyarenotrestrictedto
numericdata.
2. B.Itmixesscalardata(theCRUISE_ORDER_IDcolumn)and
aggregatedata(theCOUNTfunctionappliedtotheORDER_DATE
column).ThisisnotpossiblewithoutaGROUPBYclause.TheGROUP
BYclausecouldbeusedtotransformtheCRUISE_ORDER_IDcolumn
intoanaggregatevaluebyspecifyingGROUPBYCRUISE_ORDER_IDat
theendofthestatementbeforethesemicolonterminationcharacter.
A,C,andDareincorrect.Someaggregatefunctionscanworkwiththe
datedatatype,andCOUNTisoneofthem.
3. AandB.COUNTnumbersoccurrencesofdata.Thatdatacan
includecharacter,date,ornumericdatatypes.MINdeterminesthe
minimumvaluebutwillworkwithcharacterdatatodeterminethevalue
representingthevaluethatwouldappearfirstinanalphabeticsortingofthe
candidatevalues.
CandDareincorrect.AVGworkswithnumericdataonly;MEDIAN
workswithnumericanddate/datetimedataonly.
4. B.COUNTwilldeterminetheoccurrencesofVENDOR_IDinthe
datalisting.
A,C,andDareincorrect.SUMaddsnumbers,whichisnotdesiredhere.
MEDIANdeterminesanaveragevalue.HAVINGisnotanaggregate
function;itisaclauseoftheSELECTstatement.
5. D.Itwillexecute.Thestatementissyntacticallycorrect.
A,B,andCareincorrect.YoucanuseAVGwithaPRIMARYKEY
column.Itmightnotproduceanyusefulinformation,butit’sallowed.You
canalsouseMINwithDATEdatatypes,aswellascharacterstrings.It
doesn’tmatterifthetablehasonlyonerow;thestatementwillstillwork.
Youareallowedtouseanaggregatefunctiononzero,one,ormorerows.
6. C.TheAVGwillcomputebyignoringtheNULLvalueand
averagingtheremainingvalues,resultinginananswerof90.00.
A,B,andDareincorrect.Thereisnosyntaxerrorhere;theTO_CHAR
functioncorrectlyformatstheoutputtoincludecommasifnecessary.There
willbenoexecutionerror,either.IftheAVGfunctionincludedNULL
valuesand,say,treatedthemaszeros,thentheanswerwouldbe60.00.That
couldbeaccomplishedwiththeNVLfunction,asin
AVG(NVL(TEST_SCORE,0)).Butthat’snotincludedasanoptionhere.
7. A.Therequesttoranktheliteralnumericvalueof100000within
thesetofvaluesforPROJECT_COSTasksSQLtoestablishanumeric
rankingforthevalue100000withinthesetofrowsandindicatewherea
rowcontainingavalueof100000forPROJECT_COSTwouldfallwithin
thesortedlist,ifthereweresucharow.
B,C,andDareincorrect.Thenumericliteralvaluecannotincludea
comma.Inotherwords,withintheargumentsforRANK,thevalueof
“100,000”(withoutquotes)isnotseenasonehundredthousandbutinstead
isseenasonehundred,followedbyasecondparameterofthreezeros.Two
parameterscannotbeacceptedbyRANKunlesstherearetwo
correspondingWITHINGROUPexpressions;thereisonlyoneinthe
answerprovided.InC,theGROUPBYismisplaced.InD,'Upgrade'
representsaninvaliddatatypebecausethedatatypesofbothRANKand
ORDERBYmustmatch,andinthisexample,PROJECT_COSTisa
numericdatatype,asweseeintheaccompanyingexhibit.
8. A,B,C,andD.Allofthefunctionsmentionedignorenullvalues.
MAXinparticularisworthemphasizing—rememberthatNULLvaluessort
higherthanNOTNULLvalueswhenORDERBYsortsonacolumn
containingNULLvalues.However,whilethatistrue,onlyNOTNULL
valuesareconsideredbytheMAXfunction.Thesameistruefortheother
functionslistedinthequestion.
Alloftheanswersarecorrect.
9. BandD.TheSELECTstatementallowsanaggregatefunctiontobe
calledfromtheORDERBYclauseandtheselectlist.Ifyouspecifyan
aggregatefunctionfromwithinanORDERBYclausewhenaGROUPBY
clauseispresent,theORDERBYwillsorttheaggregaterowsthateach
representagroupofrows.
AandCareincorrect.ThereisnoHAVINGclauseinanINSERT
statement.ThereisnoexpressionlistinaDELETEstatement.
GroupDatabyUsingtheGROUPBYClause
10. A.SomemightprefertoplacethePURPOSEcolumnbeforethe
SUM(DAYS)expressionintheSELECTexpressionlist,butthatisnot
required.
B,C,andDareincorrect.Bissyntacticallyincorrectsinceyoucannot
putanaggregatefunctionwithinaGROUPBYclause.Cisincorrect
becausetheCOUNTfunctionisusedinsteadofSUM.COUNTwould
counttheoccurrencesofdata,ratherthansumupthevalues.Disjusta
randomcombinationofreservedwordsandnonsense.
11. C.Thestatementissyntacticallycorrectandwillexecute.
A,B,andDareincorrect.TheTO_CHARfunctionisascalarfunction
and,assuch,isnotsubjecttothesamerestrictionsthatanyaggregate
functionis.Forexample,theTO_CHARusedintheGROUPBYclauseis
fine.AndbyusingtheTO_CHARexpressionintheGROUPBY,itcan
alsobeusedintheSELECTexpressionlistinline1,alongwiththe
aggregatefunctionCOUNT.
12. D.Itwillsucceedanddisplayonerow.Thereasonyouknowthisis
becauseline1showsanaggregateofanaggregatewithaGROUPBY
clause.
A,B,andCareincorrect.Line1iscorrect;youareallowedtonestone
aggregatewithinoneotheraggregatefunction.YoucanuseGROUPBYto
groupanydatatype;thereisnorestrictionongroupingbycharacterdata.If
youweretouseonlyoneaggregatefunction,theresultwoulddisplayone
rowforeachuniquevalueofPURPOSE.ButbynestingCOUNTwithin
COUNT,youareaddingupallofthoserowsanddisplayingoneaggregate
answer.
IncludeorExcludeGroupedRowsbyUsingtheHAVINGClause
13. AandC.HAVINGisvalidonlyintheSELECTstatement,nottheother
SQLstatements.HAVINGmustoccuraftertheWHEREclause.
BandDareincorrect.HAVINGcannotbeusedwithoutGROUPBY,but
itisnotrequiredtofollowGROUPBY.HAVINGisnotlimitedto
expressionsidentifiedintheGROUPBYclause;instead,itislimitedto
anyvalidexpressionthataddressesthegroupsestablishedwithinthe
GROUPBYclause.Inotherwords,aggregatefunctionscanbeinvoked
fromwithinHAVING,whethertheyarecalledbyGROUPBYornot.
14. C.Thestatementissyntacticallycorrectandwillproduceaseriesofrows
ofdata.EachrowwillrepresentallvaluesforeachSHIP_IDvalueinthe
table.EachrowrepresentingeachSHIP_IDwillhaveonemaximumvalue
forDAYS.AnysetofrowsforSHIP_IDwhoseaveragevalueof
PROJECT_COSTislessthan500000willbeincluded;allotherswillbe
excluded.
A,B,andDareincorrect.Thereisnosyntaxerroronline4.The
HAVINGclauseisnotrequiredtoreferenceanythingintheGROUPBY.
HAVINGcanreferenceaggregatefunctions;itsonlylimitationisthatit
cannotdirectlyreferencecolumnsinthetableifthosecolumnshavenot
beenincludedintheGROUPBYclause.(ColumnsomittedfromGROUP
BYmaybeincorporatedincertainexpressionsinHAVINGbutcannotbe
directlyreferencedasstandalonecolumns.)Bcouldbetrueonlyifthe
WHEREclausewerepresentandfilteringrowsasBdescribes—the
HAVINGclausedoesn’tincludeorexcluderowsonanindividualbasisbut
insteadincludesorexcludesgroupsofrowsasdefinedbytheGROUPBY
clauseandtheHAVINGclause.
15. A.OneofthemostimportantaspectsofunderstandingHAVINGisto
knowwhennottouseit,andthisisagreatexampleofthat—nothingin
thisquestionsaysanythingaboutrestrictinggroupsofdata,andthatisthe
onlyreasonwhyyouwoulduseHAVING.TheWHEREclauseachieves
thetaskofensuringthatonlyrowswhereDAYSisgreaterthanthreewill
beconsidered.
B,C,andDareincorrect.Thetaskdefinedbythequestiondoesn’t
requireHAVING.Thereisnodescriptioninthequestionthatasksfor
anythingabouttheGROUPStobeexcluded.Bissyntacticallyincorrect
becauseHAVINGcanreferenceonlyaggregatefunctionsorcolumnsthat
areidentifiedintheGROUPBY,andthisdoesneither.Cissyntactically
incorrectbecauseyoucannotenclosethefinalexpressioninparenthesesby
itself.Dissyntacticallycorrectbutdoesnotproducetheanswerthatis
describedinthequestion.Itwillinsteadgroupalltherowsbyboth
PURPOSEandDAYS,whichisnotwhatwasaskedfor;suchaquerycan
potentiallyproducemanymorerowsthanjustthetwoexpected.Inother
words,insteadofgettingonegroupofrowspervalueinPURPOSE,you’ll
getafinerdivisionofdetailwitheachsetofrowscontainingaunique
combinationofbothPURPOSEandDAYS,andthatissomethingdifferent.
8
DisplayingDatafromMultipleTables
T
CERTIFICATIONOBJECTIVES
8.01DescribetheDifferentTypesofJoinsandTheirFeatures
8.02WriteSELECTStatementstoAccessDatafromMoreThanOneTable
UsingEquijoinsandNon-Equijoins
8.03JoinaTabletoItselfbyUsingaSelf-Join
8.04ViewDataThatGenerallyDoesNotMeetaJoinConditionbyUsingOuter
Joins
✓Two-MinuteDrill
Q&ASelfTest
hischapterreviewsvarioustypesofjoins.AjoinisaSELECTstatementthat
retrievesdatafromtwoormoretablesandconnectsthatdatatoreturna
mergedsetofrows.Toaccomplishthejoin,theSELECTstatementspecifies
datainonetablethatisidenticaltodatainanothertable.Tablesthatshare
commondataelementsthatsupportjoinsaresaidtoberelatedtoeachother.
ThisiswhyaSQLdatabaseiscalledarelationaldatabasemanagementsystem;
joiningtablesisthewholepurposeofSQLdatabasesystems.
Thischapterwilllookat
Equijoins,whichjoindatausingtheequalityoperator
Non-equijoins,whichjoindatausingnon-equalityoperators
Innerjoins
Outerjoins
Naturaljoins
Self-joins
We’lldiscussthesyntaxofalloftheseformsofjoinsandlookattheuseof
tablealiases,aswellasanumberofexamples.
CERTIFICATIONOBJECTIVE8.01
DescribetheDifferentTypesofJoinsandTheir
Features
ThereareseveraltypesofjoinsinSQL.Themostcommonlyusedisthe
equijoin.Lesscommonisthenon-equijoin.We’lldiscussbothinthissection.
TypesofJoins
Joinsarecharacterizedinmanyways.Onewayajoinisdefinedisintermsof
whetheritisaninnerjoinoranouterjoin.Anotherissueisthatofequijoinsand
non-equijoins.Thesedescriptionsarenotmutuallyexclusive.
Let’sbrieflylookateachtypeofjoin,afterwhichwewillanalyzeeachonein
detail,withexamples.
InnerJoinsvs.OuterJoins
TherearetwomajorcategoriesofsyntaxforcreatingjoinsinOracleSQL:
Innerjoinsconnectrowsintwoormoretablesifandonlyifthere
arematchedrowsinallthetablesbeingjoined.
Outerjoinsconnectrowsintwoormoretablesinawaythatismore
inclusive—ifdataexistsinonetablethathasnomatchingvaluesin
another,theunmatchedrowwillstillbeincludedintheoutput.
Equijoinsvs.Non-Equijoins
Separatefromtheissueofinnerandouterjoinsistheissueofequijoinversus
non-equijoin:
Equijoinsconnectdataintwoormoretablesbylookingforcommon
dataamongthetables’columns.Inotherwords,anequijoinlooksforan
exactmatchofdata.
Non-equijoinsconnectdatabylookingforrelationshipsthatdon’t
involveequality,suchas“lessthan”or“greaterthan”relationships,or
situationswheredatainonetableiswithinarangeofvaluesinanother.
Mostofthejoinswe’llworkwithwillbeequijoins,butwe’lllookatnon-
equijoinsaswell.
OtherJoins
Thereareotherjoins,suchasself,andnaturaljoins.We’llrevieweachin
subsequentsectionsofthischapter.
CERTIFICATIONOBJECTIVE8.02
UseSELECTStatementstoAccessDatafromMore
ThanOneTableUsingEquijoinsandNon-Equijoins
Thissectionlooksatvariousformsofjoins.Ajoinisaquerythatassociates
rowsofdatainonetablewithdatainanothertable.Atwo-tablejoiniscommon,
butyoucanjointhreeormoretables.Onapracticallevelaquerythatjoinstoo
manytablesintroducesperformanceandmaintenancechallenges.Buthowmany
tablesare“toomany”forajoin?MostindustryleadersIknowsayfiveisabout
thelimit,butI’veseennine-tablejoinsthatmadesenseforacertaincontextona
particularproject.
Thissectionlooksatinnerjoins,includingnaturaljoins,multitablejoins,
non-equijoins,andtheUSINGclause.
InnerJoins
Thefirsttypeofjoinwe’llreviewisaninnerjoin.Aninnerjoinreturnsa
mergedrowfromtwoormoretablesifandonlyiftherearematchedvalues
amongalljoinedtables.Foranygivenrow,ifthereisn’tamatchedvalueinall
tables,thennorowisreturned.
Forexample,let’sconnecttwotableswithaninnerjoin.Everyshipinthe
SHIPStablecanbeassignedahomeportinthePORTStable.Thisisrecordedin
thedatabasebyassigningaHOME_PORT_IDtoeachrowintheSHIPStable.
ThisiswhytheSHIPStableincludesacolumncalledHOME_PORT_ID,which
isdesignatedaFOREIGNKEY.TheFOREIGNKEYconstraintreferencesthe
PORT_IDcolumninthePORTStablesothatanyentryintheSHIPStable’s
HOME_PORT_IDwillbeidenticaltothePORT_IDforthatPORTasitalready
existsinthePORTStable’sPORT_IDcolumn.
Let’slookatoursampledatainthesetables.Firsthere’sthePORTStable:
Next,here’stheSHIPStable:
Asyoucanseefromthesedatalistings,manyshipshaveaHOME_PORT_ID
thatmatchesaPORT_IDinthePORTStable.Ifyouweretoidentifythehome
portfor,say,theCoddElegance,youwouldseethatit’sPORT_ID3,whichis
Tampa.Seeit?Didyoulookatthesampledatatoconfirmwhatwejust
described?Ifyoudid,youcomparedthevalueofHOME_PORT_IDinSHIPSto
thevalueofPORT_IDinPORTs.Thatispreciselywhatyouhavetoexplainin
SQLsyntaxsothatSQLcandothesamething.Inotherwords,tojointhese
tablestogether,wedothefollowing:
IdentifybothtablesintheFROMclause,separatedbythekeywords
INNERJOIN.
Definethecolumnfromeachtablethatisbeingusedtojointhedata
intheONcondition.
Here’sthecodeforaninnerjoin(linenumbersadded):
TheORDERBYissomethingweadded;it’snotrequiredforthejoin.
NotethatthekeywordINNERisnotrequired;wecaneliminateitandgo
withthisapproach:
ThisqueryisthesameastheINNERJOINquerywithonevariation:we
removedtheoptionalkeywordINNERfromline2.Everythingelseisthesame.
Whicheversyntaxweuse—withorwithoutthekeywordINNER—theresult
isthesame.
That’stheoutputfromourinnerjoin.
NotethatwecanaddadditionalWHEREclausecriteriatoourjoin.Let’ssay
wewantedtorestrictouroutputsothatonlytheCharlestonrowswereincluded.
Let’smodifytheINNERJOINsyntax:
Theresultwillbetolimitourresultstoonlythoseshipsthatareportedin
Charleston.
NowgobackbeforeweaddedtheWHEREclauseandlookatouroriginal
outputfromthefirstinnerjoin.Doyouseesomethingmissing?Lookforthe
CoddChampion—itisn’tintheoutput.ThereasonisbecausetheCodd
Champion,akaSHIP_ID3,hasnoassignedvalueforHOME_PORT_ID.Asa
result,thereisnomatchingrowinthePORTStable,andsinceweusedan“inner
join”format,weonlyreturnrowswheretherearematchedvaluesinthejoin
criteria.TheresultisthattheCoddChampionisomitted.Andthat’snotall—we
alsodon’tseeanyvalueforthePORT_ID4,'Miami'.Thereasonisthesame—
thereisnoshipassignedtoMiamiasahomeport.
Thisiswhyourjoinisaninnerjoin.Itonlyreturnsrecordsthathavematched
rowsinbothtables.Ifyouwanttoincludedatainouroutputfromrowsthat
aren’tnecessarilymatchedinalltables,usethe“outerjoin”format.We’lllookat
outerjoinsshortly.
OlderInnerJoinSyntax
Beforewemoveontoouterjoins,let’sreviewanoldvariationoninnerjoin
syntax.Hereitis:
(Note:Thistimeweusedtablealiases.Mostjoinsconnecttablesby
referencingcolumnswithidenticalnames.Inthisexample,thePORTStable
usesPORT_ID,andtheSHIPStableusesHOME_PORT_ID,sotablealiasesare
notrequired.However,ifSHIPShadnameditsforeignkeyPORT_ID,theuseof
tablealiasesinourexamplewouldberequired.You’lllearnmoreabouttable
aliasesinafewpages.)
Inthisform,thejoinisaccomplishedinlines2and3,wherewelistthe
joinedtablesintheFROMclauseandalsoidentifythejoincriterioninline3
withanequalsign.Inthissyntax,thereisnokeywordJOINorON.The
WHEREclausecanincludeadditionalcriteriaasanyWHEREclausemight.
Here’sanexample:
ThisistheolderformofaninnerjointhatOraclehasused,anditstillworks.
Butthatversion,whileclear,succinct,easytouse,andverypopular,is
nonethelessinconsistentwithANSIstandards.Theversionwereviewedearlier
—theversionthatusestheJOINandONkeywords—isformallypreferredandis
consistentwiththeANSIstandardforSQLjoins.
UsingTableAliases
Inourexamplessofar,we’vejoinedtablesusingcolumnsthathavehaddifferent
columnnames.Thisisn’talwaysthecase,however.Forexample,lookatthetwo
tablesinFigure8-1.
FIGURE8-1 TheEMPLOYEESandADDRESSEStables
TheEMPLOYEEStableandtheADDRESSEStablebothhaveacolumn
calledEMPLOYEE_ID.ThatcolumnisthePRIMARYKEYinEMPLOYEES,
andit’stheFOREIGNKEYinADDRESSES.Usingwhatwe’veseensofar,we
mighttrythissyntax:
There’saproblemwiththissyntax.Lookatline1.Thefirstiteminour
SELECTstatement’sselectlistisEMPLOYEE_ID,butwhichtable’s
EMPLOYEE_IDareweintendinghere?Chancesareitwon’tmakethatmuchof
adifferencewithregardtodatadisplay,sincebothcolumnscontainthesame
data—unlessweuseaformofOUTERJOIN,inwhichcasetheremaybea
difference.Regardless,that’snohelptousherebecauseSQLwillrejectthis
statementanywaybecauseoftheambiguouscolumnreference.
Onesolutionistousethefulltablenameasaprefixtothecolumn.
We’veaddedthefulltablenameinfrontofeachreferencetoEMPLOYEE_ID,
separatedbyaperiod,toclarifyourintent.Notethatweneededtomakethese
changesthreetimes—onceonline1andtwiceonline3.Thisisperfectly
acceptableandrepresentsclearandthoroughdesign.
There’sanalternative,however,andit’sthetablealias.Here’sanexampleof
thetablealiasinaction:
Firstlookatline2.AftereachtablereferencedintheFROMclause,welefta
space,followedbyanamewespecify.Thenameneedstomatchtherulesfor
namingdatabaseobjects.ForEMPLOYEESwechoseEM,andfor
ADDRESSESwechoseAD.Wecouldhavechosenanythingwithintherulesof
namingdatabaseobjects.
Sincewechosetocreatetablealiases,wewerealsoabletousethetablealias
asprefixesonlines1and3.Theresultisamoreeasilyreadablequery.
Notethescopeofthetablealiasisthestatementitself.Oncethestatementhas
completedexecuting,thetablealiasisnolongeruseful;itmustbespecified
anewinanysubsequentSQLstatements.
There’snorightorwronganswerastowhichapproachyoushouldchoose—
thefulltablenameprefixorthetablealiasprefix.Eitherway,youshould
considerthefollowingpoints:
Whenwritinganyqueryinwhichacolumnreferenceisambiguous,
youmustdosomethingtoidentifythecolumnclearlytotheSQL
statement,usingeitheratableprefixoratablealias.Otherwise,you’ll
getanerrormessageandtheSQLstatementwon’texecute.
Theuseoftableprefixesandtablealiasesisontheexam.
Thetablealiasismorecommonlyused,inmyever-so-humbleopinion.But
bothwork.
AlsonotethattablealiasescanbeusedinINSERT,UPDATE,MERGE,and
DELETEstatements,aswellasSELECT.
NaturalJoins
Sofar,manyoftheexampleswe’veseenhavejoinedtablesusingcolumnsof
differentnames,suchasHOME_PORT_IDandPORT_ID.It’smorecommon,
however,forsuchcolumnstobenamedwiththesamename.Inotherwords,
whendatabasedesignerscreatetwotablesandintendtolinkthemtogether,itis
gooddesigntogiveidenticalnamestoanycolumnsthatwillbelinked.It’snot
necessarilybaddesigntodootherwise—italldependsonreadabilityandintent.
Butthepointforourdiscussionhereonnaturaljoinsisthatyouwilloftenwork
withtablesthatshareidenticalcolumnnames,anditwillbethesecolumnsupon
whichyourjoinswillprobablybebuilt.
Anaturaljoinisaninnerjoinbydefault,butitcanalsobealeftouterjoin,
rightouterjoin,orafullouterjoin.
Forexample,reviewFigure8-1again.Aswesawintheprevioussection,the
twotablesshownbothhaveacolumncalledEMPLOYEE_ID.Thiscolumnisa
FOREIGNKEYintheADDRESSEStableandthePRIMARYKEYofthe
EMPLOYEEStable.
ThenaturaljoinapproachtellsSQLtolocateanycolumnsinthetwotables
withacommonname,andusethemtojointhetables.Here’sanexample:
NoticetheuseofthekeywordsNATURALJOINinline2.Alsonoticethat
there’snokeywordONanywhereandnowherethatweestablishthejoin
betweenthetwocommoncolumns—whichinthesetwotablesare
EMPLOYEES.EMPLOYEE_IDandADDRESSES.EMPLOYEE_ID.The
NATURALJOINdoesn’trequireanexplicitdeclarationofthesecolumns,
providedthatthecolumnnamesareidentical.
Alsonoticesomethingelse:seethereferencetoEMPLOYEE_IDinline1?
RememberthattherearetwoEMPLOYEE_IDcolumns—oneinEMPLOYEES
andoneinADDRESSES.Normallysuchareferenceinajoinwouldrequirea
tablealias.Butthenaturaljoinforbidssuchtableprefixesonjoincolumnnames.
Theirusewouldresultinasyntaxerror.However,tableprefixesareallowedon
othercolumns—butnotthejoincolumnsinanaturaljoin.
Finally,anaturaljoindefaultstoaninnerjoin,butitworkswithaleftouter
join.
Italsoworkswitharightouterjoin.
Finally,itevenworkswithafullouterjoin.
TheSELECTlistmayreferenceacolumnthatiscommontobothtablesbut
notqualifiedwithatablenameprefix.Ifthishappens,whichtabledoesthe
valuecomefrom?Theanswerdependsonwhichtypeofnaturaljoinisused:
Ifaninnerjoinoraleftouterjoinisused,theleftmosttableisthe
sourceofthevalue.
Ifarightouterjoinisused,therightmosttableisthesource.
Ifafullouterjoinisused,bothtables’rowswillbecombinedtoformthe
output,sothevaluewillbetakenfromthecombinationofbothtables.
USING
ThekeywordUSINGissimilartothenaturaljoin,inthesensethatitsuse
dependsonthepresenceofidenticallynamedcolumnsintheJOIN.Like
NATURAL,USINGcanbeusedwithbothinnerandouterjoins.
Let’slookatanexample.FollowingisalistingofdataintheEMPLOYEES
table:
NextisalistingoftheADDRESStable:
Here’saquerythatjoinsthetwotableswithaUSINGclausethatspecifies
theEMPLOYEE_IDcolumn:
Notethesyntaxhere—we’reusingaLEFTJOINontwotablesthatsharethe
samecolumnnameEMPLOYEE_ID.Aswe’vealreadyseen,thisisavariation
ofanouterjoin.ThekeywordOUTERisoptionalandomittedfromthis
example.
NoticethekeywordUSINGonline3.USINGisfollowedbyacolumnname
enclosedinparentheses.Notablenameprefixisallowedbeforethecolumn
name—nothere,andnotelsewhereinthestatementsuchasline1.
Sincethetablenameprefixisnotallowedinline1,howdoyouknowwhich
table’sEMPLOYEE_IDisintendedintheselectlistinline1?Theansweristhat
itmaybeoneortheother,dependingontherow.Inanouterjoin,whichisalso
anequijoin,thevaluesinbothtablesforEMPLOYEE_IDeitherwillbeidentical
toeachotherorwillincludeoneNULLvalueinoneofthetwotables'columns.
Theoutput,therefore,willshowavalueifanyispresentineithertable;
otherwise,aNULLvaluewilldisplay.Thereasonthisworksisthatthevalues
willneverconflictinanouterjoin.Thejoincondition,bydefinition,willreject
conflictingvaluesandexcluderowswithconflictingvaluesforthejoin
condition.Theresultisthatifyouincludethejoincolumnintheselectlist—
suchastheEMPLOYEE_IDcolumninline1oftheexample—thentheoutput
ofanouterjoin’s“joincolumn”specifiedwiththeUSINGclausewillshowa
valuewhereveravalueispresentineithertable’sjoincolumn;otherwise,itwill
displayaNULLvalue.
Let’sextendourexample.
Thisexampleassumesthepresenceofanadditionalcolumn
(OFFICE_NAME)inbothtablesandperformsthejoinusingthecombinationof
EMPLOYEE_IDandOFFICE_NAME.
MultitableJoins
Sofarwe’velookedonlyatjoinsthatconnecttwotables.Butjoinscanconnect
two,three,ormoretables.Forexample,seeFigure8-2.
FIGURE8-2 PORTS,SHIPS,andSHIP_CABINStables
ThesethreetablescanbejoinedinaSELECTstatementlikethis:
NoticethesyntaxforthisSELECTstatementthatjoinsthreetables:
TheFROMkeywordappearsonce.
Afterline2,whentheoriginaltwo-tablejoincompletes,line3opens
withthekeywordJOIN.
Line3continueswiththethirdtablename,followedbyatablealias,
followedbytheexplicitlydefinedjoincriteria.
Line3intheprecedingSELECTstatementcanberepeatedandeditedas
requiredinordertojoinadditionaltablesbeyondthesethree.
Non-Equijoins
Sofar,allthejoinswe’veseenhavebeenequijoins.Inotherwords,thejoins
haveusedcolumnscontainingcommondatavaluesamongthejointablesand
havejoinedrowsinthetablesbasedonfindingequalvaluesinthejoincolumns.
Non-equijoinsrelateonerowtoanotherbywayofnon-equalcomparisons,
suchascomparisonsofgreaterorlesservalueorperhapscomparisonsthatlook
forarangeofvalues.Suchjoinsareunusualbutimportant.Andtheyarepartof
thecertificationexamcriteria.
Let’slookatanexample.SeeFigure8-3.
FIGURE8-3 TheSCORESandGRADINGtables
NotethatthesetablesdonothaveaFOREIGNKEYrelationship.Aswe’ve
alreadynoted,thatisnotrequiredtojointables.
Let’slookatsomedatainthesetables.First,here’sSCORES:
Next,GRADING:
TheideahereisthattheSCOREStablelistssomeactualscoresonexams,and
theGRADINGtablecontainsinformationaboutgradingcriteria.
IntheSCOREStable,wehavearowofdatashowingthatthetestidentified
withaSCORE_IDvalueof1receivedascoreof95.Accordingtothe
GRADINGtable,thatshouldbeagradeofA.
Let’screateaSELECTstatementthatjoinsthesetablestodetermineeachtest
score’sgrades.
Notethesyntaxofthisjoin.Online2wehaveatypicalJOINsyntax,butonline
3,insteadoftheequijoin,weconnectthetwotablesbycomparingthevaluein
theTEST_SCOREcolumnofSCORESandseewhetherit’sBETWEENthe
valuesfortheGRADINGtable’sSCORE_MINandSCORE_MAXcolumns.
Hereistheoutput:
Thisisanexampleofanon-equijoin.ThesyntaxoftheONconditionina
non-equijoinissimilartothesyntaxfortheWHEREclauseintheSELECT,in
thatyoucanusecomparisonexpressions,thegreater-thanorless-thanoperators,
SQLfunctions,andBooleanoperatorstoconnectaseriesofcomparisons.
CERTIFICATIONOBJECTIVE8.03
JoinaTabletoItselfbyUsingaSelf-Join
Aself-joinisatablethatisjoinedtoitself.Aself-joinisbasedonthevalueina
columnofatableasitcomparestothevalueswithinanothercolumn—oftena
differentcolumn—inthesametable.Theself-joinreturnsoneormorerows
fromatablethataremergedwithotherrowsinthesametablebasedonthejoin
criteria.
(Note:Syntacticallyyoucanjoinacolumntoitselfinthesametable,as
opposedtoadifferentcolumninthesametable.Itdoesn’tdomuchlogically,but
thesyntaxwillexecute.)
Self-joinscanuseallthesamevariationsonjoincriteriathatanyothertable
joincanuse.Inotherwords,self-joinscanbeinnerjoinsorouterjoins,equijoins
ornon-equijoins,andsoon.
Let’swriteaself-joiningSELECTstatement.Forstarters,seeFigure8-4.The
POSITIONStablelistsjobtitleswithinourCoddCruisesorganization.
FIGURE8-4 ThePOSITIONStable
Here’salistingofsomeofthecolumnsinthePOSITIONStable:
NotethecolumnREPORTS_TO.Itindicatesthat,forexample,CrewChief
reportstothepositionofPOSITION_ID2,whichisDirector.
Self-ReferencingForeignKeys
ThePOSITIONStableissupportedwithaforeignkeythatpointstoitself,which
iscreatedwiththisSQLstatement:
Notethataforeignkeyisadvisedbutnotrequiredinordertoperformthe
self-join.
Self-JoinSyntax
Nowthatwehaveourtable,ourdata,andouroptionalforeignkeyconstraint,
let’slookataquerythatwillconnectallofthistogether.Tocreatetheself-join,
weneedtodothefollowing:
IdentifythetabletwiceintheFROMclause.
Defineourjoincriteria—whichinthiscasewillbeanOUTERjoin
sothatwecanincludethehighest-levelpositionthatdoesn’thavea
REPORTS_TOvalue—andthereforeisn’t“matched”withanything.
Applyatablealiastoallappropriatereferences,beingcarefultojoin
theREPORTS_TOcolumninthefirsttablereferencetothe
POSITION_IDcolumninthesecondtablereference.
Here’sthequery:
Andhere’stheoutput:
Besureyouhaveasolidcommandofthejoinsyntaxforallformsof
joins.ThatshouldincludeNATURAL,USING,andthevariousinner
joinsandouterjoins.Allformsarefairgameontheexam.
Theresultisalistingofpositionsandthesupervisoreachreportsto.Allof
thisdataisfoundintheonetablePOSITIONS.YoucanseethatintheSELECT
statement’soutput,boththePOSITIONcolumnandtheBOSScolumncontain
datathatisfoundinthePOSITIONcolumnofthePOSITIONStable.Butthe
self-joinreturnsameaningfuloutputdisplaywithacolumnaliasonthesecond
referencetoPOSITION.
CERTIFICATIONOBJECTIVE8.04
ViewDataThatGenerallyDoesNotMeetaJoin
ConditionbyUsingOuterJoins
Anouterjoinisajointhatreturnsallthesamedataasaninnerjoinbutalso
returnsadditionalrows,namely,thoserowsthatdon’thavematchesinallthe
tablesthatarejoinedtogether.Therearethreetypesofouterjoins—LEFT,
RIGHT,andFULL.Eachisdescribedhere.
LEFTOUTERJOIN
Toseeonetypeofouterjoininaction,let’scontinuewithourexampleand
modifyourqueryjustalittletomakeitanouterjoin.
NoticetheadditionofthekeywordsLEFTOUTERinline2.Thefollowingare
therowsreturnedbythiscode:
BychangingourFROMclausefromaJOINtoaLEFTOUTERJOIN,we
havechangedthequeryfromaninnerjointoanouterjoin.(Rememberthat
JOINdefaultstoINNERJOIN.)Andnoticethatouroutputnowincludesdata
forSHIP_ID3.AlsonoticethattherowforSHIP_ID3showsnovaluefor
PORT_NAME,whichiscorrect—noHOME_PORT_IDvaluewasassignedto
it.
AlsonotethattheOUTERkeywordisoptional.Inotherwords,thisisjustas
good:
Inthisexample,OUTERisomitted,buttheeffectisthesame.
RIGHTOUTERJOIN
Butwestilldon’tseeourPORT_NAMEofMiami.So,let’schangeourquery
again.
Notethatwe’vereplacedthekeywordLEFTwiththekeywordRIGHTinline
2.Hereistheresult:
We’venowincludedtherowforMiami.ButwelostourrowwiththeCodd
Champion.That’sbecausethisRIGHTOUTERJOINreturnsunmatchedrows
fromtherightsideofthejoin,whichinthisexampleisthePORTStable(see
line2oftheSELECTstatement).ButwetookawayourLEFTOUTERJOINto
SHIPS.
Asbefore,OUTERisoptional.RIGHTJOINwilldothesamethingas
RIGHTOUTERJOIN.
FULLOUTERJOIN
IfwewanttocombinetheeffectsofaRIGHTOUTERJOINandLEFTOUTER
JOIN,wecanusetheFULLOUTERJOIN.
Line2showstheuseofthekeywordsFULLOUTERJOIN,andtheoutputisas
follows:
Andherewehaveallofourrows,mergedwherematched,butallreturned
regardless.Aswithallotherouterjointypes,thekeywordOUTERisoptional.
FortheRecord:OracleOuterJoinSyntax:(+)
TheLEFT,RIGHT,andFULLOUTERsyntaxusagesareANSIstandardand
featuredonOracle’sexam.Inaddition,OracleSQLoffersanalternativeouter
joinsyntaxthatemploysaplussignenclosedinparentheses.Thissyntaxisused
quiteabitandlongpredatestheANSIstandardsyntaxinOracleSQL;however,
theouterjoinplusoperatorisnotontheexam.
Let’slookatthesyntaxanyway.Ifyou’reaveteranOracleprofessional,you
mightthinkthisbookisincompletewithoutatleastamentionoftheouterjoin
plusoperator.Let’srestatetheearlierLEFTOUTERJOINexamplewiththeplus
operator.
Notethespecialcharactersattheendofline3.Theplussigninparentheses
definesthisqueryasaleftouterjoin.Itisaleftouterjoin,buttheplussigngoes
ontherightside;IthinkoftheplusoperatorasaddingfakeNULLrowsonthe
rightsidetoenabletheotherwiseunmatchedrowsonthelefttobeincludedin
thereturnedrowset.
Tocreatearightouterjoininthissyntax,yousimplymovetheplussignto
theleftside.
Youcan’treallydoafullouterjoinusingthissyntax,butyoucanusetheset
operatorUNIONtocombinealeftouterjoinandarightouterjointoachievethe
sameresult.
(Note:SetoperatorsareaddressedindetailinChapter11.)
So,that’stheplusoperator.ButitwasneverANSIstandard.Furthermore,
OracleCorporationformallyrecommendsthatyouavoidusingitfromnowon
andstickwiththekeywordsINNERJOIN,OUTERJOIN,andalltheother
relatedkeywords.Inspecialsituationstheplusoperatornowhassome
restrictionsandlimitationsthatdon’tpertaintotheINNERJOINandOUTER
JOINkeywords.
Butyouwon’tbetestedonit—you’llbetestedonthenewerLEFT,RIGHT,
andFULLOUTERJOINformats.
CERTIFICATIONSUMMARY
ThereareseveralwaystojointablesinSQL.Theinnerjoincomparescommon
valuesinrowsoftwoormoretablesandreturnsdatafromarowonlyifthatrow
hasamatchinthejoinedtable.Theouterjoinwillshowarowasoutputwhether
thatrowhasajoinedrowinanothertableornot.
Somejoinsareequijoins,whichmeansthattheyjoinrowsinonetablewith
rowsinanothertablebylookingforvaluesinbothrowsthatareequaltoeach
other.Non-equijoinslookforvaluesthathavesomerelationshipotherthan
equality,suchasgreaterthan,lessthan,orinbetweenarangeofvalues.
Primarykeyandforeignkeyrelationshipshelpprotecttheintegrityofdatain
columnsthatareintendedtojointables,buttheirpresenceisnotrequiredin
ordertocreateasuccessfuljoin.
Whentwoormoretablesarejoined,anyoftheircolumnsmaybeincludedin
theSELECTstatement’sselectlist.However,asyntaxerrormayresultifthetwo
tableshavecolumnsthatsharethesamename.Insuchsituations,thetablename
canbeplacedinfrontofthecolumnnameasaprefixtoeliminateanyambiguity.
Asanalternative,thetablecanbeassignedanaliasnamethatwilllastonlylong
enoughfortheSELECTstatementtoexecute.Thealiascanthenbeusedasa
prefixinfrontofanycolumnname,anditmayberequiredinfrontofthe
otherwiseambiguouscolumnnames.
Anaturaljoindoesnotspecifywhichcolumnsarebeingusedtoconnecttwo
ormoretablestogether,relyinginsteadontheassumptionthatthecolumnsto
formthejoinhavethesamenameaseachother.Naturaljoinsareinnerjoinsby
default.
TheUSINGkeywordcandosomethingsimilartotheNATURALkeyword.
USINGcannamethecommoncolumnonetime,andtheSELECTstatementwill
completethejoinbasedonthatcolumn,aswellaseliminatingtheneedfora
tableprefixoraliasinfrontofthekeycolumn.
Multitablejoinscanbeperformedwithanyjointype.
Non-equijoinsusecomparisonoperators,Booleanlogic,oranythingelseto
establishcomparisonlogicbetweentwoormoretablesinajoin.
Aself-joinoccurswhenatableisjoinedtoitself.Typicallyonecolumninthe
tableisjoinedtoadifferentcolumninthesametable.
✓TWO-MINUTEDRILL
DescribetheDifferentTypesofJoinsandTheirFeatures
Theinnerjoincomparesarowinonetabletorowsinanothertable
andreturnsthefirsttable’srowonlyifamatchingrowinthesecond
tableisfound.
Theouterjoincomparesarowinonetabletorowsinanothertable
andreturnsthefirsttable’srow,aswellasdatafrommatchingrowsin
thesecondtablewhereveramatchisfound.
Theequijoinidentifiesaparticularcolumninonetable’srows,
relatesthatcolumntoanothertable’srows,andlooksforequalvaluesin
ordertojoinpairsofrowstogether.
Thenon-equijoindiffersfromtheequijoininthatitdoesn’tlookfor
exactmatchesbutinsteadlooksforrelativematches:greaterthan,less
than,orarange,suchasonetable’svaluethatisbetweentwovaluesin
thesecondtable.
Theself-joinconnectsatabletoitself.
WriteSELECTStatementstoAccessDatafromMoreThanOne
TableUsingEquijoinsandNon-Equijoins
AjoinisanySELECTstatementthatselectsfromtwoormore
tables.
Ajoinwillconnectrowsinonetablewithrowsinanothertable.
ThetablealiasexistsonlyforthedurationoftheSQLstatementin
whichitisdeclared.
Tablealiasesarenecessarytoeliminateambiguityinreferringto
columnsofthesamenameinajoin.
Thenaturaljoindoesnotnametheconnectingcolumnbutassumes
thattwoormoretableshavecolumnswithidenticalnamesandthatthese
areintendedtobetheconnecting,orjoining,columns.
TheUSINGkeywordcanempoweraninner,outer,orotherjointo
connectbasedonasetofcommonlynamedcolumns,inmuchthesame
fashionasanaturaljoin.
Joinscanconnecttwo,three,ormoretables.
JoinaTabletoItselfbyUsingaSelf-Join
Theself-joinconnectsatabletoitself.
Self-joincriteriacomparesacolumninatablewiththesameor
anothercolumninthesametable.
Self-joinscanalsobereferredtoasrecursivejoins.
Self-joinscanbewrittenasequijoins,non-equijoins,innerjoins,
andouterjoins.
ViewDataThatGenerallyDoesNotMeetaJoinConditionby
UsingOuterJoins
Theouterjoincomparesrowsintwotablesandreturnsoutput
whetherthereisamatchingrowornot.
Theleftouterjoinshowsalltherowsinonetableandonlythe
matchingrowsinthesecond;therightouterjoindoesthesamethingin
reverse.
Thefullouterjoinshowsallrowsinbothtablesonewayorthe
other—eitherasamatchedrowsetorasastandalonerow.
SELFTEST
Thefollowingquestionswillhelpyoumeasureyourunderstandingofthe
materialpresentedinthischapter.Chooseallthecorrectanswersforeach
question.
DescribetheDifferentTypesofJoinsandTheirFeatures
1.Aninnerjoinqueriesfromtwotables(lookingatvaluesincolumns
andoptionallyusingexpressionsthatreferencecolumns)andcomparesthe
resultingvaluesinonesetofrowswiththeresultingvaluesinanothersetof
rows,lookingfor:
A.Valuesthatmatch
B.Valuesthatmayormaynotmatch
C.Valuesinthefirstsetthataregreaterthanvaluesinthesecond
set
D.Valuesinthefirstsetthatarelessthanvaluesinthesecondset
2.WhichofthefollowingsymbolsismostlikelytobeusedinaSELECT
statementusinganon-equijoin?
A.!=
B.<>
C.<=
D.Noneoftheabove
3.Equijoinslookfor:
A.Exactdatamatches
B.Rangesofdatamatches
C.Comparisonsusinganycomparisonoperatorprovidedthatthe
resultingcorrelationsoccurinbothtables
D.Noneoftheabove
WriteSELECTStatementstoAccessDatafromMoreThanOne
TableUsingEquijoinsandNon-Equijoins
4.ReviewthisSQLstatement:
Whichoneofthefollowingkeywordsinthisstatementisoptional?
A.JOIN
B.INNER
C.ON
D.Allarerequired
5.ReviewtheINVOICESandVENDORStables.
NextreviewthefollowingSQLstatement:
WhichofthefollowingstatementsistruefortheSQLstatement?
A.Itwillexecutesuccessfully.
B.ItwillfailwithasyntaxerrorbecausethereisnoONclause.
C.Itwillfailwithasyntaxerroronline1becauseVENDOR_IDis
ambiguous.
D.Itwillfailwithasyntaxerroronline3becauseofthe
parenthesesaroundVENDOR_ID.
6.Reviewtheillustrationfromquestion5andthenreviewthefollowing
SQLstatement:
Whatkindofjoinisthis?(Choosetwo.)
A.INNER
B.OUTER
C.NATURAL
D.Equijoin
7.Atablealias:(Choosetwo.)
A.Renamesatableinthedatabasesothatfuturejoinscanusethe
newname.
B.Isthesamethingasadatabaseobjectsynonym.
C.ExistsonlyfortheSQLstatementthatdeclaredit.
D.Canbeusedtoclearupambiguityinthequery.
8.ReviewthePOSITIONS,EMPLOYEES,andPAY_HISTORYtables.
ReviewthefollowingSQLstatement:
WhichofthefollowingistruefortheSQLstatement?(Choosetwo.)
A.Itwillfailbecausetherearenotablealiases.
B.Itwillexecutesuccessfully.
C.Itisanouterjoin.
D.Itconnectsthreetables.
9.Reviewtheillustrationfromquestion8andthenreviewthefollowing
SQLstatement:
WhichofthefollowingstatementsaccuratelydescribetheSQL
statement?(Choosetwo.)
A.Itcontainsasyntaxerroronline3.
B.Itisaninnerjoin.
C.Itisanon-equijoin.
D.Itcontainsasyntaxerroronline2andshouldhaveanadditional
keywordwiththeJOINkeyword.
10.Howmanytablescanbejoinedinaquery?
A.Onlytwo
B.Asmanyasyoulike,providedtheyareallconstrainedwith
PRIMARYKEYandFOREIGNKEYconstraintstoensurethatthe
joinconditionwillwork
C.One,two,three,ormore
D.Nomorethanseven
11.Youhavetwotables.OnetableiscalledCUSTOMERS.Anotheriscalled
PURCHASES,anditrecordsalistofcustomertransactions.Yourgoalisto
createaSELECTstatementthatwillshowallcustomersbylastnamein
alphabeticalorder,alongwithanypurchasestheymayhavemadeinthe
pasttwoweeks,asrecordedinthePURCHASEStable.It’spossiblethat
manycustomershavemadenopurchasesinthepasttwoweeks,butyou
stillwantthemincludedintheoutput.Bothtablescontainacolumncalled
CUSTOMER_ID.WhichofthefollowingwillbetrueoftheSELECT
statementyou’llneedtocreate?(Choosetwo.)
A.Itwillbeaninnerjoin.
B.Itwillbeanouterjoin.
C.Itwillbeacross-join.
D.Itwillbeanequijoin.
JoinaTabletoItselfbyUsingaSelf-Join
12.Aself-joinis:(Choosetwo.)
A.ASELECTstatementthatspecifiesonetableonceintheFROM
clause
B.ASELECTstatementthatspecifiesonetabletwiceintheFROM
clause
C.ASELECTstatementthatjoinsatabletoitselfbyconnectinga
columninthetabletoadifferentcolumninthesametable
D.ASELECTstatementthatusestheSELFJOINkeywords
13.Reviewtheillustrationfromquestion8.Whichofthefollowingisavalid
self-joinstatement?(Chooseallthatapply.)
ViewDataThatGenerallyDoesNotMeetaJoinConditionby
UsingOuterJoins
14.Reviewtheillustrationfromquestion5.Whichofthefollowingisa
syntacticallycorrectouterjoinquery?(Choosetwo.)
15.ThedifferencebetweenanINNERandanOUTERjoinis:
A.TheINNERjoinrelatesatabletoitself;theOUTERjoinrelates
atabletoothertables.
B.TheINNERjoindisplaysrowsthatmatchinalljoinedtables;
theOUTERjoinshowsdatathatdoesn’tnecessarilymatch.
C.TheOUTERjoinrelatesatabletotablesinotheruseraccounts;
theINNERdoesnot.
D.TheINNERrunsondatainsidethetable;theOUTERrunson
dataoutsideofthetable.
SELFTESTANSWERS
DescribetheDifferentTypesofJoinsandTheirFeatures
1. A.Aninnerjoinlooksforvaluesthatmatch.
B,C,andDareincorrect.Aninnerjoinignoresrowswherevaluesdon’t
match;anouterjoinmayreturnsomeofthoserows.Anon-equijoinmay
leveragematchesdefinedbygreater-thanorless-thancomparisonoperators.
2. C.Thenon-equijoinperformscomparisonsusingrelationshipsthat
don’tinvolveequalitybutwherethereisstillsomesortofmatchingcriteria,
suchaslessthan,greaterthan,orvalueswithinarange.
A,B,andDareincorrect.Thenot-equalsigns!=and<>arenotwhatis
intendedwithanon-equijoin.Remember,thesearenot“not-equal”joins
butrather“non-equijoins.”
3. A.Equijoinslookforexactdatamatches.
B,C,andDareincorrect.Non-equijoinsuserangesofmatches.
Equijoinsdon’tuseanycomparisonoperator;theyuseonlyequalsigns.
WriteSELECTStatementstoAccessDatafromMoreThanOne
TableUsingEquijoinsandNon-equijoins
4. B.INNERisoptional.WhencreatinganINNERJOIN,youcanjust
usethetermJOINwithoutthereservedwordINNER,andthejoinwillbe
assumedtobeINNER.
A,C,andDareincorrect.JOINisrequired.SoisON.
5. A.Itwillexecutesuccessfully.
B,C,andDareincorrect.EventhoughthisisaJOIN,thepresenceof
USINGeliminatestheneedforanONclause.WithUSING,atablealiasis
notrequired,norisoneallowedintheselectlist.Theparenthesesaround
VENDOR_IDonline3arefine.
6. AandD.TheINNERkeywordisoptional,butthatiswhatthisisby
default.It’salsoanequijoin,inthatthetwotablesarejoinedbydefining
valuesthatmatchequallyintheVENDOR_IDcolumn—thisisspecifiedby
useoftheUSINGclauseonline3,whichisequivalenttoON
VENDORS.VENDOR_ID=INVOICES.VENDOR_ID.
BandCareincorrect.ItisnotanOUTERjoin;ifitwere,theOUTER
keywordmustbespecified.ItisnotaNATURALjoinforthesamereason
—ifitwere,theNATURALkeywordmustbespecified.
7. CandD.Thetablealiasgoesawayafterthequeryisover.Itcanbe
usedtoclearupconfusioninthesyntaxofastatementbyaddingitasa
prefixatthebeginningofanycolumnthatsharesthesamenamewith
anothercolumninajoinedtable.
AandBareincorrect.Thetablealiasisnotthesameasadatabaseobject
synonym.ThetablealiassurvivesonlywithintheSQLstatementthatcalls
it;afterthat,itgoesaway.
8. BandD.Itwillexecutesuccessfully,anditconnectsthreetables.
AandCareincorrect.Itdoesnotrequiretablealiases.Itisaninnerjoin,
notanouterjoin.
9. BandC.Thequeryisaninnerjoin.ThekeywordINNERis
optionalandomittedinthisinstance.Thequeryisanon-equijoin,wherein
theONconditioncomparesthePAY_HISTORYtable’sSALARYcolumn
ofvaluestotheMAX_SALARYandMIN_SALARYofthePOSITIONS
table.
AandDareincorrect.Thestatementcontainsnosyntaxerrors.
10. C.Youcanjoinasmanytablesasyouwant.Rememberthatjoiningone
tabletoitselfisaself-join.
A,B,andDareincorrect.Youarenotlimitedtotwotables,norareyou
requiredtoonlyjointablesthathaveconstraintsonthem.Youarenot
limitedtoseventables.
11. BandD.TheSELECTmustbeanouterjointoincludeallrecordsinthe
CUSTOMERStablewhetherornottheyhaveacorrespondingrowinthe
PURCHASEStableinthelasttwoweeks.Also,sincebothtablescontaina
CUSTOMER_ID,you’lluseanequijointolocateexactmatchesbetween
thetwotables.
AandCareincorrect.Thequerycannotbeaninnerjoin,becauseifit
were,itwouldonlyshowcustomerswhohavemadepurchasesinthepast
twoweeks.Itcannotbeacross-join,whichisaCartesianproduct,because
allthatwoulddoisconnecteverycustomerwitheverypurchasewithout
anyregardforwhetherthecustomeractuallymadethepurchaseor
someoneelsedid.
JoinaTabletoItselfbyUsingaSelf-Join
12. BandC.Inaself-join,thetablenameisrepeatedtwiceintheFROM
clause,andperhapsmorethantwice.Aself-joinismostcommonlyusedto
joinacolumninatabletoadifferentcolumnwithinthesametable.
AandDareincorrect.Theself-joinsyntaxnamesthesamenametwice
—ormore—intheFROMclause.ThereisnosyntaxinOracleSQL
involvingthekeywordsSELFJOIN.
13. A,C,andD.AllofthesearevalidSQLstatements.Thefirstisatypical
innerjoin.CspellsouttheoptionalwordINNERforthesameresultasA.
Ddefinesavalidouterjoinandisalsoaself-join.
Bisincorrect.ThereisnokeywordSELF,asinSELFJOIN.
ViewDataThatGenerallyDoesNotMeetaJoinConditionby
UsingOuterJoins
14. AandC.TheLEFTJOIN...ONsyntaxiscorrect.SoistheRIGHT
OUTERJOIN...ONsyntax.RememberthatthekeywordOUTERis
optional.
BandDareincorrect.TheOUTER...ONsyntaxisincorrect;thereis
noOUTERbyitself.TheFULLOUTER...ONsyntaxisalsoincorrect.
InneitheroftheseistherequiredkeywordJOINpresent.
15. B.INNERshowsdataonlywhenthere’sarow-to-rowmatchbetween
tables,whereasOUTERcanshowdataforrowsinonetablethatdon’t
matchtherowsinthejoinedtable.
A,C,andDareincorrect.Adescribesajointhat“relatesatableto
itself.”Thatisaself-join,notaninnerjoin.Thefactthatatablemayor
maynotexistinotheruseraccountsisirrelevant;youcanjointables
whetherornottheyareinthesameuseraccount.OUTERdoesnotrunon
dataoutsideofthetable;alljoinsreferencedatawithineachoftheir
respectivereferencedtables.
9
T
UsingSubqueriestoSolveQueries
CERTIFICATIONOBJECTIVES
9.01DefineSubqueries
9.02DescribetheTypesofProblemsSubqueriesCanSolve
9.03DescribetheTypesofSubqueries
9.04QueryDataUsingCorrelatedSubqueries
9.05UpdateandDeleteRowsUsingCorrelatedSubqueries
9.06UsetheEXISTSandNOTEXISTSOperators
9.07UsetheWITHClause
9.08WriteSingle-RowandMultiple-RowSubqueries
✓Two-MinuteDrill
Q&ASelfTest
hischapterlooksatsubqueries.AsubqueryisaSELECTstatementwithina
largerSQLstatement.Thesubqueryconceptbuildsonmanyofthefeatures
we’vealreadylookedatandcombinestheminawaythatmakesSQLquite
powerful.However,subqueriesintroducesomecomplexitiesthatareimportant
tomaster.Theyareanimportantpartofthecertificationexam.
CERTIFICATIONOBJECTIVE9.01
DefineSubqueries
Asyou’vealreadyseen,aqueryisaSELECTstatement.Asubqueryisa
SELECTstatementthatexistswithinanotherSQLstatement—sortofa“sub”
SQLstatement.SQLstatementsthatacceptsubqueriesincludeSELECT,
INSERT,UPDATE,MERGE,andDELETE.SubqueriesarenotlimitedtoDML;
theycanalsobeusedinaCREATETABLEorCREATEVIEWstatement.
Mostsubqueriesaresyntacticallyautonomous,meaningyoucouldexecute
thesubquerysuccessfullyonitsown,separatefromtheparentquery.However,
somesubqueriesarenotstandalonequeriesbutinsteadarecorrelatedtothe
largerparentqueryinawaythattiesthetwoqueriestogether.
Asubquerymayreturnoneormorecolumnsinoneormorerows.One
commonuseofasubqueryistoreturndatathatisofusetotheparentquery,
suchascriteriawithinaWHEREclause,forexample.ASQLstatementthat
includesasubqueryaspartofitscodeisconsideredtheparenttothesubquery.
AparentSQLstatementmayincludeoneormoresubqueriesinitssyntax.
Subqueriesmayhavetheirownsubqueries.Inotherwords,aSQLstatement
mayhaveasubquery,whichinturnmayhaveanothersubquerywithinit,andon
andon.ThestartingSQLstatementinsuchasituationisthetop-levelquery,and
anysubsequentsubquerythatcontainsasubquerywithinitisaparenttoits
subquery.Aparentquerymayalsobereferredtoasanouterquery.
AnyvalidSELECTstatementcanqualifyasasubqueryofanotherstatement.
SELECTstatementsthatcanbemadeintoasubqueryincludethosethatretrieve
datafromoneormoretablesorviews;thosethatusecomplexexpressionsor
scalaroraggregatefunctions;andthosethatincludetheWHEREandGROUP
BYclauses,HAVINGclauses,joins,oranyoftheotherfeaturesavailableto
SELECTstatements.
Asubqueryisnotrestrictedtoretrievingdatafromthesametableortablesas
theparentquery.Infact,subqueriesoftenretrievedatafromtablesotherthan
thosespecifiedintheparentquery.Thetablesofthesubqueryarenotrequiredto
haveanykeyrelationshiporjoinorotherlogicalrelationshiptothetablesofthe
parentquery.Sucharelationshipmayexistbetweentheparentandthesubquery
—butitisnotrequired.
Subqueriescanbeusedtosolveavarietyofproblems.
CreatingcomplexmultistagequeriesSubqueriescanfindanswers
toquestionsandthenusethoseanswerstoasknewquestions.Inother
words,inasingleSELECTstatement’sWHEREclause,asubquerycan
finddatabasedonsomecriteria,thenusethatanswertoidentifya
secondaryanswer,andfinallypassthatanswerontotheparentquery,all
inthesameSQLstatement.Thiscanberepeatedwithinmultiplestages
inwaysthatastandaloneparentquerycannotdowithouttheuseofa
subquery.
CreatingpopulatedtablesAsubquerycanbeincorporatedintoa
CREATETABLEstatementtoquicklycreateandpopulateatablefrom
anexistingdatasourceatthetimethetableiscreated.
ManipulatinglargedatasetsSubqueriescanbeincorporatedinto
anINSERTorUPDATEstatementtomovelargeamountsofdataand
insertorupdatemanyrowsofinformationbymovingdatafromone
sourceintoanotherinasingleSQLstatement.
CreatingnamedviewsAsubqueryisusedtocreateviewobjectsat
thetimeofcreation.
DefiningdynamicviewsAsubquerycanbeusedtoreplaceatable
referenceinaFROMclause.Thisformofsubquery,alsoknownasan
inlineview,isdiscussedinChapter10.
DefiningdynamicexpressionswithscalarsubqueriesThereisa
formofsubquerythatwillreturnonlyonecolumn’sworthofdatainone
row;thatis,itwillreturnasinglevalue.Suchasubquerycanbeusedin
almostanyplaceinaSQLstatementwhereanexpressioncanbeused.
Asyoucansee,subqueriesarepowerful,candoawidevarietyoftasks,and
canoccurinalmostanySQLstatementinalmostanyclause.Thischapterwill
exploremanyexamplesoftheseusagesofsubqueries.
CERTIFICATIONOBJECTIVE9.02
DescribetheTypesofProblemsSubqueriesCanSolve
Subqueriesareanidealsolutiontoseveralcommondatachallenges.Sometimes
asubqueryistheonlysolution.Subqueriesareidealforfindingdatathatisatthe
topofsomescale,isaboveaverage,isbelowaverage,orsomethingcomparable.
Youmightfirstuseaquerythatinvolvesanaggregatefunctionandthenusethe
resultofthatquerytofindrecordsthatcorrelatetotheresult.
Let’slookatanexamplewithsomecode.ConsideratableSALES_DATA.
Let’saddrecordstothetableforsalesrepresentativesshowingsalesbyyear
fortheyears2016through2018.
Youaretaskedwithlistinganyandallsalesrepswhosesalesfor2016
surpassedthesalesteam’saveragesalesfortheprioryearof2017.
Onewaytodothisistofirstgettheaveragesalesfor2017.
Next,usetheresultstolookforsalesrepsin2018whosoldatahigherlevel.
That’sonewaytogettheanswer.Butasubquerywouldhavefoundthe
answerinasinglequery.
Youcanusesubqueriesformanypurposes.
Topopulateatableusingzerotomoresourcerowswithinan
INSERTstatementordovirtuallythesamethingwithinaCREATE
TABLEstatement.
TocreateaVIEW.
TospecifyvaluesassignedtoanUPDATEstatement.
Toperformcomparisonsagainstaggregatedresults.
Tocreateinlineviews:asubqueryinaFROMclauseofaSELECT
statement.Thereisnolimittohowmanysubqueriesdeepyoucango
withinaninlineview;youcanhavesubqueriesofsubqueriesof
subqueriesinthiscontext.
Asanalternativetoexpressions.
Tonestotherqueries.AsubqueryintheWHEREclauseofa
SELECTstatementisknownasanestedsubquerybecauseitisnested
withinthelargerquery(theSELECTstatementitself).Youcannestupto
255subqueriesinthisway.
TohaveacorrelatedsubquerywithinaSELECT,UPDATE,or
DELETE.
Theremainderofthischapterwillexploretheseandotherexamplesof
subqueries.
CERTIFICATIONOBJECTIVE9.03
DescribetheTypesofSubqueries
Therearemanytypesofsubqueries.Herearethemajortypes:
Single-rowsubqueriesAsingle-rowsubqueryreturnsasinglerow’s
worthofdatainitsresult.Single-rowsubqueriesmayincludemultiple
columns’worthofdataorasinglecolumn’sworthofdata.
Multiple-rowsubqueriesAmultiple-rowsubqueryreturnszero,
one,ormorerowsinitsresult.Itisnotguaranteedtoreturnmultiple
rows,butitmaydoso,andthustheparentqueryshouldbestructuredto
receivemultiplerowsjustincaseinordertoavoidthepossibilityofan
executionerror.Forexample,aparentquerymayuseanequalsignto
compareitsownreturnvalueswiththereturnedvaluesofasubquery.An
equalsignperformsacomparisontoasinglevalue,sointhecaseofthe
subquery,onlyonerowshouldbereturned.Butifthesubqueryreturns
multiplerows,theparentquery(andthereforethequeryasawhole)will
failwithanexecutionerror—itdependsonwhetherthesubqueryactually
returnsmultiplerows.Itmightnot,butitcertainlycanatanytime.
Therefore,theparentquerytoamultirowsubquerytypicallyusesa
comparisonoperator(suchasIN)thatallowsformultiplerowsofvalues
tobereturned.
Multiple-columnsubqueriesMultiple-columnsubqueriesreturn
morethanonecolumnintheirresult.Thisrequirestheparentqueryto
receivemultiplecolumnsfromthesubqueryandinvolvesspecialsyntax
considerations,aswejustdiscussed.Amultiple-columnsubquerycanbe
eitherasingle-rowsubqueryoramultiple-rowsubquery.
CorrelatedsubqueriesAcorrelatedsubqueryisasubquerythat
specifiescolumnsthatbelongtotablesthatarealsoreferencedbythe
parentquery.Inamultilevelseriesofsubqueries,subquerieswithin
subqueries,andsoon,thecorrelatedparentquerycanbeanynumberof
levelshigherthanthesubquery.Thecorrelationinvolvesmorethanthe
subquerymerelyaccessingthesametablesandcolumnsthroughitsown
directcalltothetable;rather,thecorrelatedsubqueryperformsrow-by-
rowanalysisincooperationwiththeparentquery,accessingdataand
referencingthatdatainitsownexpressionsinordertocoordinaterow
processingwiththecorrelatedparent.Correlatedsubqueriescanexistin
SELECT,UPDATE,andDELETEstatements.Acorrelatedsubquery
mayalsobeasingle-row,multiple-row,ormultiple-columnsubquery.
ScalarsubqueriesIfasingle-rowsubqueryconsistsofonlyone
column’sworthofoutput,thenitisknownasascalarsubquery.Scalar
subqueriescanbeusedinalmostanylocationthatanexpressioncanbe
used,whichisnottrueforotherformsofsubqueries.Ascalarsubquery
canalsobecorrelated.
Notethatthedifferenttypesofsubqueriesaren’tmutuallyexclusive.A
singletypeofsubquerymayfallintomultiplecategoriesofsubqueries
describedinthischapter.
Asyoucansee,therearemanytypesofsubqueries.Let’sbeginlookingat
themindetail.
CERTIFICATIONOBJECTIVE9.04
QueryDataUsingCorrelatedSubqueries
Acorrelatedsubqueryisaquerythatisintegratedwithaparentquery.
Correlatedsubqueriesincludereferencestoelementsofaparentquery,andthus,
theydonotexistasstand-alonequeriesliketheexamplesyou’veseensofar.Up
tonow,anyofthesubqueriesyou’veseencouldhavebeenexecutedontheir
own.That’snotthecasewithcorrelatedsubqueries.
Let’stakealookatanexampleusingtheSHIP_CABINStable.InChapter7,
yousawtheSHIP_CABINStableinFigure7-2,alongwithadatalisting
showingthe12rowsofinformationinoursampletable.Eachrowin
SHIP_CABINSshowsthatacabinisofROOM_STYLESuiteorStateroom.
Eachindividualroom’svalueforSQ_FTisshown,andthevaluesarenotallthe
same.
Ourcurrentchallengeistocreateasinglequerythatlistsallthecabinsinthe
shipwhosesize—asmeasuredbytheSQ_FTcolumn—islargerthantheaverage
cabinforitsROOM_STYLE.
Inotherwords,weneedto
IdentifytheaveragesquarefootageforeachROOM_STYLEin
SHIP_CABINS
Useaveragesquarefootagevaluetodisplayeachshipcabinwhose
SQ_FTisgreaterthantheaverageforitsROOM_STYLE
Withoutacorrelatedsubquery,we’dberequiredtocreateseparatequeriesto
getthisinformation.Wewouldneedthefollowing:
Onequerytogettheaverages
Anotherquery—orqueries—tocomparetheindividualaveragesfor
eachROOM_STYLE
Butwithacorrelatedsubquery,wecandoitallatonce.Here’stheSQL(line
numbersadded):
Notethesubquerythatstartsonline3andcontinuesthrough(andincludes)line
5.Inparticular,notethesecondROOM_STYLEattheendofline5.Seethe
tablealiasofA?Thatisareferencetoacolumnoftheparentquery—notthe
subquery.Seeline2toconfirmthattheparentquery’stableisaliasedwiththeA
prefix.
Thisisthecorrelationinthiscorrelatedsubquery.Thisqueryisnotexecuting
asastand-alonequeryandthenpassingbackitsresultlikenoncorrelated
subqueriesdo.Instead,thecorrelatedsubqueryisexecutingonceforeachvalue
thattheparentqueryfindsforeachrow,passingthevalueforthe
ROOM_STYLEcolumnintothesubquery,anddeterminingtheaveragesquare
footageforthatparticularROOM_STYLE.Finally,itusestheresultofthat
queryinline3todeterminewhethertherowintheparentqueryisgreaterthan
theaverageofSQ_FTfortheROOM_TYPEornot.
Here’stheoutput:
Onewaytovalidatetheseresultswouldbetogetalistofeach
ROOM_STYLEanditsaveragevalueforSQ_FT.Here’saquerytocalculate
thatinformation:
ThisoutputconfirmstheaverageSQ_FTforeachROOM_STYLE,whichis
informationwecouldusetogobackandconfirmthatourcorrelatedsubquery
displayedonlythoseroomnumberswhoseindividualSQ_FTvaluesarehigher
thantheaveragefortheappropriateROOM_STYLE.Andtheyare.
CorrelatedsubqueriescanexistinSELECT,UPDATE,andDELETE
statements.Theyare“correlated”bywayofacolumnreferencefromtheparent
querywithinthesubquery.
Atablealiasisnotnecessarilyrequiredinthesubqueryifnocolumnname
conflictexists.Inourexample,therewassuchaconflict,sowewererequiredto
useatablealias.Butthat’snotnecessarilyrequiredinanycorrelatedsubquery.
Wecouldhavejustreferencedanycolumnfromtheparentquery,andaslongas
thereisn’tanidenticallynamedcolumninthesubquery,notablealiasis
required.
It’simportanttonotethatcorrelatedsubqueriesmayintroduceperformance
degradationintoaquery.Theprocessofcorrelatingrowsfromoneormore
subquerieswiththeouter,orparent,queryorqueriesmayconsumeasignificant
amountofprocessingtime.However,sometimesacorrelatedsubquerycan
accomplishtasksthatnootherformofquerymayaccomplish.
CERTIFICATIONOBJECTIVE9.05
UpdateandDeleteRowsUsingCorrelatedSubqueries
Let’slookattheuseofcorrelatedsubqueriesinanUPDATEstatementanda
DELETEstatement.
UPDATEwithaCorrelatedSubquery
AnUPDATEstatementcanhaveacorrelatedsubqueryintheseplaces:
IntheSETclause
IntheWHEREclause
Aswesawinanearliersection,thecorrelatedsubquerywillrequiresome
waytoidentifyanexpressionasbeingfromtheparentquery.Themostcommon
waytodothisistoassignatablealiastothetablenameintheUPDATEand
thenusethatsametablealiaswithinthecorrelatedsubquery.
Let’slookatanexample.First,reviewtheINVOICEStableshowninFigure
9-1.
FIGURE9-1 TheINVOICEStable
Ourtaskistogobacktoourhistoricalinvoicesandgivea10percent
discounttowhomeverplacedoursinglebiggestorderfortheirrespective
quarter.Sointhefirstquarter,weneedtofindthesinglebiggestinvoiceand
determinea10percentdiscountonthatinvoice,thendothesamethingforthe
secondquarter,andsoon.
Toaccomplishthisfeat,we’llneedtomodifytheinvoicesothatwechange
thevalueintheINVOICEStable’sTERMS_OF_DISCOUNTcolumntothe
string'10PCT'—andonlyfortheappropriateinvoicerecord.Andwe’llneedto
dothefollowing:
IdentifytherowwiththehighestvalueforTOTAL_PRICEforany
givenquarter,whichwecanidentifyusingtheTO_CHARformatmask
QontheORDER_DATEcolumn
UpdateaninvoiceonlyifithasthehighestTOTAL_PRICEforthe
quarter
Here’sanUPDATEstatementthatdoesthetrick:
Noticethefollowingitemsintheprecedingquery:
WechoosetocreateatablealiascalledINV,whichisdeclaredatthe
endofline1andreferencedinline6.
Thesubquerystartsinline3andrunsthroughline6.
Thecorrelationoccursinlines5and6,wherethesubquery’s
referencetoINVOICE_DATEiscomparedtotheparentquery’s
INV.INVOICE_DATE.Thecomparisonisperformedbyconvertingboth
datestothe'RRRR-Q'format,whichmeanstheyearandquartersothat,
forexample,thedate'31-MAY-17'wouldconvertto'2017-2'.
Online3isthecomparisonbetweentheparentqueryandthe
subquery,whichisanequalsign,indicatingthattheparentqueryis
expectingasingle-rowsubquery.Giventhatthesubqueryisreturningthe
aggregatevalueMAXforTOTAL_PRICEandnoGROUPBYis
involvedinthesubquery,thenthesubquerywillindeedreturnnomore
thanonerow.
HereisanexampleofacorrelatedsubqueryusedintheSETclauseofan
UPDATEstatement:
Intheprecedingcode,wedothefollowing:
Lines5through7WelookforrecordsinthePORTStablethatany
shipcallsitshomeport,bywayoftheHOME_PORT_IDcolumn.We
usethekeywordEXISTS,whichwewillreviewlaterinthischapter.
EXISTStellsusquitesimplywhetherthereareanyrowsreturnedbythe
subqueryatall.Inotherwords,we’resimplyaskingwhetherthereare
anyrowsinSHIPSthatcontainavalueforHOME_PORT_IDthat
correspondstoagivenrowintheparentUPDATEstatement’sPORTS
table.Ifyes,thesubqueryreturnsaTRUE.
Lines2through4Then,ifwefindsuchPORTS,weupdatetheir
capacitytoequalthetotalnumberofshipsintheSHIPStablecurrently
callingthatporthome;thisisaccomplishedinthecorrelatedsubquery
thatcountsthenumberofrecordsinSHIPSthatsharethesame
HOME_PORT_IDvaluewiththeparentUPDATE.
NotethatthisexampleofanUPDATEusestwocorrelatedsubqueries;bothof
thesubqueriesarecorrelatedbecausebothincludetheP.PORT_IDreferencein
theirWHEREclauses,onlines4and7.
NotethatasubqueryoftheformSELECTCOUNT(*)FROMTABLEwill
alwaysbeasingle-rowsubquery.Ifnorowsarefound,thesubqueryreturnsa
singlerowwithavalueofzero.Ifmultiplerowsexistinthetable,thesubquery
returnsasinglerowwithavaluerepresentingthenumberofrowsfound.This
istrueforqueriesusingtheCOUNTfunctionbutnotforqueriesusingother
aggregates,suchasAVG,MIN,MAX,orSUM.ThisisuniquetoCOUNT.
Nextwe’lllookathowtousecorrelatedsubqueriesinaDELETEstatement.
DELETEwithaCorrelatedSubquery
TheDELETEstatementcanbeusedwithacorrelatedsubqueryintheWHERE
clausetodeterminewhichrowstodeletefromagiventable.Thesyntaxis
similartothecorrelatedsubquerysyntaxforSELECTandUPDATEstatements.
CorrelatedsubqueriesarenotlimitedtotheSELECTstatement.They
canalsobeusedinUPDATEorDELETEstatements.
Let’stakealookataDELETEstatementfortheSHIP_CABINStable.This
DELETEwillremovethosecabinswiththesmallestbalconysquarefootagefor
eachROOM_TYPEandROOM_STYLE.Inotherwords,foragiven
ROOM_TYPEofSuiteandaROOM_STYLEofOceanView,we’llremovethe
rowforthecabinwiththesmallestbalcony,andthenwe’llmoveonanddothe
samefortheStateroomwithOceanViewthathasthesmallestbalcony,andso
on.Here’sthequery:
Noticethatinthisexample,thecorrelationinvolvestwocolumns,inlines5
and6.
CERTIFICATIONOBJECTIVE9.06
UsetheEXISTSandNOTEXISTSOperators
TheEXISTSkeywordtestsfortheexistenceofanyrowsinasubquery.Ifno
rowsarefound,theanswerisFALSE.Otherwise,thesubqueryreturnsTRUE.
NOTEXISTSreversestheresults.
Let’slookatanexample.ThefollowingquerylooksforPORTSthathaveany
sortofrecordatallintheSHIPStablewithaHOME_PORT_IDvaluethat
matchesanyofthePORT_IDvalues.Here’sthequery:
EXISTSdoesnotcompareanythingtothesubquery.Thereisno
“expressionequalsexpression”formatwithEXISTS.Itssyntaxis
simple:thekeywordsWHEREEXISTSandthesubquery.Nothingmore.
NotethekeywordEXISTSonline3.Theentiresubqueryisexecuted,even
thoughEXISTSneedonlyknowwhetherthesubqueryreturnsanyrows,so
bewareusingEXISTSwithsubqueriesthatreturnlargenumbersofrows.
It’sworthnotingthatthissortofqueryissometimesreferredtoasasemijoin,
whichisaSELECTstatementthatusestheEXISTSkeywordtocomparerows
inatablewithrowsinanothertable.
CERTIFICATIONOBJECTIVE9.07
UsetheWITHClause
YoucanusethekeywordWITHtoassignanametoasubqueryblock.Oncethe
nameisassigned,youcanreferencethenamefromelsewhereinthequery.
WITHisconsideredaclauseoftheSELECTstatement.
Let’slookatanexample.We’lluseWITHtodeclaretwodifferent
subqueries.We’llnameonePORT_BOOKINGSandtheother
DENSEST_PORT(seethefollowinglisting,lines2and8)andtheninvokeboth
ofthembynameinaSELECTstatement(lines12through14).Here’sthecode:
Theseriesofoneormoresubqueryblocksdefinedbeforethe
SELECTstatementisreferredtoasthesubqueryfactoringclause.WITH
candefineonesubqueryfactoringclause;itmustbedefinedbeforethe
SELECTstatement.
NotethatneitherPORT_BOOKINGSnorDENSEST_PORTisadatabase
object.Theyarethenamesofqueriesthatexistsolelywithinthis
WITH/SELECTstatement.
Alsonotethesubqueriesonlines3through6andonlines9through10.Also
notethattheonlysemicolonisattheendoftheentirestatement,notattheend
ofanyindividualSQLstatementwithintheoverallWITHstatement.
Internally,OracleSQLtreatsanamedquerywithintheWITHclauseasa
temporarytableorasaninlineview.(WeexaminedinlineviewsinChapter8.)
TheWITHclausecanbeusedinthetop-levelqueryofaSELECTstatement
andinmany(butnotall)subqueriesoftheSELECTstatement,suchasshownin
line14.IfyouuseWITHtonameasubquery,thatnameisn’trecognizedwithin
thesubqueryitselfbutisrecognizedinmosteveryotherlocationintheoverall
query.Inotherwords,considerline8,wherewenameandspecifythesubquery
DENSEST_PORT.Wecouldn’treferencethenameDENSEST_PORTinlines9
through10,butwecanreferencethenameeverywhereelse.
CERTIFICATIONOBJECTIVE9.08
WriteSingle-RowandMultiple-RowSubqueries
Inthissection,we’llcreatesubqueriesthatreturnoneexpressionineithera
single-rowormultiple-rowanswer.
Single-RowSubqueries
Asingle-rowsubqueryisasubquerythatreturnsonerow.Forexample,let’ssay
you’retaskedwiththejobofidentifyingthenamesofemployeesofourfictional
company,CoddCruiseLines.You’retaskedtolookforemployeeswhoare
assignedtothesameshipasanemployeenamedAlSmith.SeeFigure9-2for
theentity-relationshipdiagram(ERD)ofthetablewe’llworkwith.
FIGURE9-2 TheEMPLOYEEStable
Youcouldperformthistaskintwosteps.Step1istofindtheSHIP_IDvalue
oftheshiponwhichAlSmithisassigned,likethis:
Step2istousetheSHIP_IDvaluetolocateotheremployeesonthesame
ship.
There’sananswer—intwosteps.Notbad.
Butasubquerycouldhavediscoveredtheanswerinonestep,likethis:
Thisonequeryachievesthesamethingastheprevioustwoqueries.Notewhat
we’vedonehere—we’veeditedthesecondoftheprevioustwoqueriesby
replacingSHIP_ID=1withSHIP_ID=andthesubqueryonlines3through6.
We’veliterallyplacedthefirstoftheprevioustwoquerieswithinthesecondof
thosetwoqueries.
Thereisaninherentriskwiththisparticularsyntax.Noticetheequalsignon
line3.That’stheparentquery’swayofsayingthatitisexpectingone—andonly
one—rowfromthesubquery.Inotherwords,itisexpectingthisquerytobea
single-rowsubquery.
Here’stheproblem:whathappensifthatsubqueryreturnsmorethanone
row?WhatifthereismorethanoneAlSmithintheEMPLOYEEStable?
Let’sfindout—let’strysearchingforemployeeswhoworkonthesameship
asanyonewhoselastnameisSmith,regardlessoffirstname.
Noticethetextoftheerrormessage:“single-rowsubqueryreturnsmorethan
onerow.”Inotherwords,weapparentlydohavemorethanonepersoninthe
EMPLOYEEStablewithalastnameofSmith.
Thiscreatesaproblem.Ourqueryhasanequalsignonline3,andthat’swhy
oursubqueryisexpectedtobeasingle-rowquery.
Thereareafewalternativeswecouldtakehere.Ifwewanttoretainthe
integrityofthesingle-rowsubquery,thenweneedtoeditthatsubquerytoensure
thatitwillreturnone,andonlyone,row.
Therearemanywaystoguaranteeaone-rowresponse.Forexample,a
subquerythatusesaWHEREcriterionbasedonaprimarykeyorsomeother
uniquevaluewillreturnonlyonevalue,likethis:
Ofcourse,thatassumesyouknowthevaluefortheappropriate
EMPLOYEE_ID.
Anotherexamplewouldbeasubquerythatreturnsanaggregatefunction
withoutaGROUPBYclause,likethis:
Aggregatefunctionsalwaysreturnasinglerow,aslongasnoGROUPBYis
involved.Logically,though,thisisn’talwaysanoptiontosolvetheparticular
businesschallengeyoumightbefacing.
StillanotherexamplewouldbeasubquerythatusestheROWNUMfeatureto
limitthenumberofrowsintheresponse,likethis:
RememberthattheROWNUMpseudocolumnassignsrownumberstothe
query’soutput(beforeprocessingtheORDERBYclause)andcanbeeffectivein
limitingoutput.Inthisexample,itensuresthatwereceiveonlyaone-row
response.Butagain,thismaynotlogicallysupporttheoriginalintentofyour
query.
Themoralofthestoryistousethesingle-rowsubquerywhenyouknowthe
resultwillbeonerow.Otherwise,you’dbetterusesomethingotherthanthe
single-rowsubqueryformatoryou’llruntheriskofanexecutionerror.
Also,ifthesubqueryreturnsnorowsatall,thennoerrorwillresult.The
querywillexecute,butthesubquerywillreturnaNULLvaluetotheparent
query.
KeepinmindthatafailedexecutionattemptinSQLisn’tnecessarilyabad
thing.Sometimesit’sagoodideatocreateasingle-rowsubqueryinsituations
whereyouwanttodeliberatelyfailtheparentqueryifmultiplerowsarefound
inthesubquery.Thiscanbeausefulfeaturethatcanhelpflagbigger
problemselsewhereinthedatabase,suchasthefailureofauniqueidentifying
mechanismlikeaprimarykey,orsomeotherlogicalerrorbeyondyour
immediateSQLcode.WhenyourSQLstatementsareembeddedwithinsome
otherprogramsthathavethecapabilitiesofhandlingfailures,orexceptionsas
theyarecalledinlanguageslikeOraclePL/SQLandJava,youcanprogram
thosesystemstotakeevasiveorcorrectiveactionsandaddresswhatmightbea
muchbiggerproblem.
YoucanusemultiplesubquerieswithinasingleWHEREclause.Here’sa
SQLstatementwithmorethanonesubquery:
Noticeintheexamplewhatishappeningonline6.Inthatportionofthe
query,wearecomparingaparentcolumnnamedSSNtothesubquerycolumn
calledSOCIAL_NUMBER.Subqueriesdonothavetosharethesamecolumn
namewiththeparentquery.Theonlythingthatisrequiredisthatthedatatypes
forSSNandSOCIAL_NUMBERbecomparable.Ifthedatatypesmatchup,the
comparisonisvalid.
Table9-1liststhecomparisonconditionsavailableforusewithsingle-row
subqueries.
TABLE9-1 Single-RowSubqueryComparisonConditions
ThekeywordINdeservesspecialattention.WhenthekeywordINisused,the
parentqueryseesthesubqueryaspotentiallyamultiple-rowsubquery,which
we’lldiscussinthenextsection.
So,single-rowsubqueriesareusefulforperformingmultistepquerieswhen
youcanguaranteethatthesubquerywillreturnasinglerow.Butwhenyour
subqueryreturnsmultiplerows,youneedtotakeadifferentapproach.Let’slook
atthatnext.
Multiple-RowSubqueries
Amultiple-rowsubquerymayreturnmorethanonerowofanswerstotheparent
query.Forexample,wesawinourearlierexamplethatthereisapparentlymore
thanoneemployeeinourEMPLOYEEStablewithalastnameofSmith.
Becauseofthatfact,wecouldn’tusethesingle-rowqueryformatwitha
subquerythatsearchedforrowswithaLAST_NAMEvalueofSmith.Butwe
canusethatsubqueryinthemultiple-rowformat,asshownhere:
Nowourqueryisaskingtolisttheemployeeswhoworkonashipwith
anyonenamedSmithandthenlisttheSmithsthemselves.Butforourpurposes,
noteline3,wherewereplacedtheequalsignwiththereservedwordIN.Asa
resultofthatonechange,wenolongergetanerrormessagebutinsteadgetthis:
RememberfromourdiscussionoftheWHEREclauseandthesetof
comparisonoperatorsthatthekeywordINallowstheWHEREclauseto
compareonevaluetoasetofvalues.Thesamedynamicisatworkhere.IN
allowsthesubquerytoreturnmultiplerows.ThepresenceofthekeywordIN
directstheparentquerytoallowthesubquerytobeamultiple-rowsubquery.
Notethatthesubquerydoesn’thavetoreturnmultiplerows.Thepresenceof
INsimplyallowsforthepossibility.
Noticewedidn’tgetanexecutionerrorthistime.Instead,weobtainedoutput
andcannowseethatwehaveanemployeenamedSmithassignedtoSHIP_ID1
andanotheronSHIP_ID3.OurSELECTstatementreturnsrowsshowing
employeesoneithership.
Table9-2liststhecomparisonoperatorsthatcanbeusedwithamultirow
subquery.
TABLE9-2 MultirowSubqueryComparisonConditions
Youcanusegreater-thanorless-thancomparisonoperatorswithsingle-row
subqueriesbutnotmultirowsubqueriesunlessthoseoperatorsarecombined
withALL,ANY,orSOME.
ThekeywordsshowninTable9-2indicatethattheparentqueryisexpecting
thesubquerytobeamultirowsubquery.Amultirowsubquerycanreturn
anywherefromzerotomultiplerowsofanswers.
NOTINandNULLValuesintheSubquery
Considerthefollowing:
Sofar,sogood.Butlet’saddaNULLtotheKIDtableandtrytheNOTIN
again.
Sowhathappened?WhenaNULLisintroducedintotherowsetreturnedby
thesubquery,ano-winsituationresults.Howcanyouknowdefinitivelywhether
theparentqueryHOUSEvalueisnotequaltoanyofthevaluesreturnedbythe
subquerywhenthosevaluesincludeNULL?Remember,NULLisnot“blank.”It
is“Idon’tknow”—itisanunknownvalue.SoistheparentHOUSEvaluenot
equaltoavaluethatisn’tknownbythesubquery?Whocouldtell?Itis
impossibletotell,so“norowsselected.”
TheNOTINcomparisonoperatorwillalwaysruntheriskofthissituationif
thesubqueryincludesNULLvalues.OneapproachistoexcludeNULLvaluesin
thesubquerywhenusingNOTIN.
Theresultmakessense—butonlyifanyNULLvaluesareexplicitlyomitted
fromthesubquery.
CERTIFICATIONSUMMARY
AsubqueryisaSELECTstatementthatexistswithinalargerSQLstatement.
SubqueriescanbeincludedinaSELECT,INSERT,UPDATE,MERGE,or
DELETEstatement.SubqueriescanalsobeusedinaCREATETABLE
statement.
SubqueriescanbeusedinWHEREclausesofSELECT,UPDATE,and
DELETEstatements.TheycanbeusedintheUPDATE...SETclauseandthe
INSERTlistofvalues.Somesubqueriesmaybeabletosubstituteforany
expressionalmostanywhereanexpressionisaccepted,includingtheSELECT
listofaSELECTstatement.
Subqueriescanperformmultiple-stepqueriesinasingleSQLstatement.
Theycanbeusedtoreferencelookupinformationfromagivenquery.Theycan
populateatableatthetimeofcreationinaCREATETABLEstatement.They
areusedtocreateviews.
Therearemanytypesofsubqueries,includingsingle-row,multiple-row,
multiple-column,scalar,andcorrelated.
Single-rowsubqueriesreturnonerowofdatatotheparentquery.Multiple-
rowsubqueriescanreturnanywherefromzerotoonetomorethanonerow.
Multiple-columnsubqueriesarecomparedtorowsintheparentqueryusing
multiplecolumnsatonce.Scalarsubqueriesreturnonerowandonecolumn’s
worthofdataatalltimes.Correlatedsubqueriescontainconditionsinthe
subquerythatconnectrowsofdatawithrowsintheparentquery,muchlikea
joinmightdo.
✓TWO-MINUTEDRILL
DefineSubqueries
AsubqueryisaSELECTstatementcontainedwithinaSQL
statement.
TheouterSQLstatementiscalledtheparent.Theoutermostlevelis
thetoplevel.
Atop-levelSQLstatementcontainingasubquerycanbea
SELECT,INSERT,UPDATE,orDELETE,orelseaCREATETABLE
orCREATEVIEW.
Subqueriescanbenestedwithinothersubqueries.
Manysubqueriescouldfunctionasstandalonequeries.Someare
correlated,meaningthattheycontainreferencesthattiethemintotheir
parentqueries.
DescribetheTypesofProblemsSubqueriesCanSolve
Subqueriesareidealforqueriesbasedonaggregateresultsofother
queries.
Subqueriesareusedtocreateviewobjects.
YoucanuseasubqueryinaSELECT,UPDATE,orDELETE
statement.
Subqueriescanbeusedasalternativestoexpressions.
DescribetheTypesofSubqueries
Asingle-rowsubqueryreturnsonerowofdatatotheparentquery.
Amultiple-rowsubquerymayreturnmorethanonerowofdatato
theparentquery.
Multiple-columnsubqueriesreturntwoormorecolumns’worthof
dataatoncetotheparentquery,whichmusttestforallthecolumnsat
once.
Correlatedsubqueriesusedatafromaparentquerytodetermine
theirownresult.
Scalarsubqueriesalwaysreturnonevalue,representedinone
columnofonerow,everytime.
Themultiple-columnsubquerymaybeofthesingle-rowor
multiple-rowtypeofsubquery.
Acorrelatedsubquerymightbeasingle-row,multiple-row,or
multiple-columnsubquery.
QueryDataUsingCorrelatedSubqueries
Correlatedsubqueriesusedatafromtheparentinsubquery
predicatestodeterminewhatdatatoreturntotheparentquery.
Correlatedsubqueriesmaypresentsomeperformancedegradation;
however,theycanperformtasksthatcouldnototherwisebe
accomplishedinasinglequery.
UpdateandDeleteRowsUsingCorrelatedSubqueries
TheUPDATEandDELETEstatementscanusecorrelated
subqueries.
TheUPDATEstatementcanusecorrelatedsubqueriesintheSETor
WHEREclause.
TheDELETEstatementcanusecorrelatedsubqueriesinthe
WHEREclause.
UsetheEXISTSandNOTEXISTSOperators
TheEXISTSoperatorcanbeusedbyaparentquerytotesta
subqueryanddeterminewhetheritreturnsanyrowsatall.
NOTEXISTSreversesthefindingsofEXISTS.
UsetheWITHClause
TheWITHclausecandynamicallynameasubquerysothatthe
SELECTstatementfollowingtheWITHclausecanreferencethat
subquerybyname,treatingitasadynamictableinrealtime.
AnysubquerynamesassignedwithintheWITHclauseareonly
goodforthatstatement;theyarenotstoredinthedatabase.
WriteSingle-RowandMultiple-RowSubqueries
Theresultsofasingle-rowsubquerycanbecomparedtoan
expressionwithintheparentusingascalarcomparisonoperator,suchas
anequalsignoragreater-thanorless-thansign.
Thecolumnnamesarenotrequiredtomatchinsuchacomparison,
butthedatatypesmustmatchsothattheparentquerycancompare
columnsofanynametosubquerycolumnsofanyname,providedthe
datatypesmatch.
Multiple-rowsubqueriesarecompareddifferentlytotheparent
querythansingle-rowsubqueries.
Multiple-rowsubqueriesarecomparedtoexpressionsintheparent
querybyusingmultiple-rowcomparisonconditions,suchasIN,ANY,
orALL,incombinationwithsingle-rowcomparisonoperatorssuchas<,
>,=,!=.
Amultiple-rowsubquerycannotbecomparedtoaparentquery
expressionusingasingle-rowcomparisonoperator;ifsuchanattemptis
made,thestatementwillfailwithanexecutionerrormessage.
SELFTEST
Thefollowingquestionswillhelpyoumeasureyourunderstandingofthe
materialpresentedinthischapter.Chooseonecorrectanswerforeachquestion
unlessotherwisedirected.
DefineSubqueries
1.Whichofthefollowingformsofsubqueryneverreturnsmorethanone
row?
A.Scalar
B.Correlated
C.Multiple-column
D.Noneoftheabove
DescribetheTypesofProblemsSubqueriesCanSolve
2.Whichofthefollowingproblemscanbesolvedwithasubquery?
(Choosethetwobestanswers.)
A.Youaretaskedwithdeterminingtheminimumsalesforevery
divisioninamultinationalcorporation.
B.Youaretaskedwithdeterminingwhichdivisionsinacorporation
earnedsaleslastyearthatwerelessthantheaveragesalesforall
divisionsintheprioryear.
C.Youaretaskedwithcreatingaview.
D.Youaretaskedwithcreatingasequence.
DescribetheTypesofSubqueries
3.Whichofthefollowingstatementsaretrue?(Choosetwo.)
A.Asingle-rowsubquerycanalsobeamultiple-rowsubquery.
B.Asingle-rowsubquerycanalsobeamultiple-columnsubquery.
C.Ascalarsubquerycanalsobeamultiple-columnsubquery.
D.Acorrelatedsubquerycanalsobeasingle-rowsubquery.
4.Whichsubqueryincludesreferencestotheparentqueryandthus
cannotexecuteasastandalonequery?(Choosethebestanswer.)
A.Ascalarsubquery
B.Acorrelatedsubquery
C.Amultiple-columnsubquery
D.Areferentialsubquery
5.Whichofthefollowingcanasubquerybeusedin?(Chooseallthat
apply.)
A.AnINSERTstatement’sSELECT
B.AGRANTstatement
C.AWHEREclauseinaSELECTstatement
D.Aninlineview
6.Aninlineviewisaformofasubquery.
A.True
B.False
QueryDataUsingCorrelatedSubqueries
7.ReviewthisWORK_HISTORYtable.
Yourtaskistocreateaquerythatwilllist—foreachship—allofthe
EMPLOYEE_IDvaluesforalltheemployeeswhohavetheshortestwork
historyfortheirship.Inotherwords,iftherearetwoships,youwanttolist
alltheemployeesassignedtothefirstshipwhohavetheshortestwork
history,alltheemployeesassignedtothesecondshipwhohavethe
shortestworkhistory,andsoon.Whichofthefollowingquerieswill
accomplishthistask?(Choosetwo.)
8.ReviewthePORTSandSHIPStables.
Yourteamistaskedwiththejobofcreatingalistoftheshipswiththe
leastcapacityineachport.Inotherwords,eachshiphasahomeport.For
eachportthatisahomeporttoships,whichofeachport’sshipshasthe
leastcapacity?Yourteamproducesthefollowingqueryinanswertothis
task:
WhichofthefollowingstatementsistrueaboutthisSQLstatement?
A.Thestatementwillfailwithasyntaxerrorbecauseofthe
subqueryonlines1through3.
B.Thestatementwillfailwithanexecutionerrorbecauseofthe
subqueryonlines1through3.
C.Thestatementwillexecutebutwillreturnmeaningless
information.
D.Thestatementwillexecutesuccessfullyasintended.
9.Acorrelatedsubquery:
A.MaybeusedinaSELECTbutnotanUPDATE
B.Cannotbeexecutedasastandalonequery
C.Mustuseatablealiaswhenreferencingacolumnintheouter
query
D.Alloftheabove
UpdateandDeleteRowsUsingCorrelatedSubqueries
10.Whichofthefollowingcanacorrelatedsubquerybeusedin?(Choose
three.)
A.TheSETclauseofanUPDATEstatement
B.TheWHEREclauseofanUPDATEstatement
C.TheWHEREclauseofaDELETEstatement
D.TheFROMclauseofaDELETEstatement
11.Reviewtheillustrationfromquestion8andthefollowingSQLcode:
ThePORTStablehas15rows.TheSHIPStablehas20rows.Eachrow
inPORTShasauniquevalueforPORT_ID.EachPORT_IDvalueis
representedintheHOME_PORT_IDcolumnofatleastonerowofthe
SHIPStable.WhatcanbesaidofthisUPDATEstatement?
A.ThevalueforCAPACITYwillincreaseonceforeachofthe15
rowsinthePORTStable.
B.ThevalueforCAPACITYwillincreaseby20foreachofthe15
rowsinthePORTStable.
C.ThevalueforCAPACITYwillnotincrease.
D.Thestatementwillfailtoexecutebecauseofanerrorinthe
syntax.
UsetheEXISTSandNOTEXISTSOperators
12.AnothernameforanEXISTSqueryis:
A.Demijoin
B.Multiple-columnsubquery
C.Cross-join
D.Semijoin
13.Reviewtheillustrationfromquestion8andthefollowingSQLcode:
ThecodeisattemptingtodeleteanyrowinthePORTStablethatisnot
ahomeportforanyshipintheSHIPStable,asindicatedbythe
HOME_PORT_IDcolumn.Inotherwords,onlykeepthePORTSrows
thatarecurrentlytheHOME_PORT_IDvalueforashipintheSHIPS
table;getridofallotherPORTrows.That’stheintentoftheSQL
statement.WhatwillresultfromanattempttoexecutetheprecedingSQL
statement?
A.Itwillfailbecauseofasyntaxerroronline2.
B.Itwillfailbecauseofasyntaxerroronline4.
C.Itwillfailbecauseofanexecutionerrorinthesubquery.
D.Itwillexecutesuccessfullyandperformasintended.
UsetheWITHClause
14.TheWITHclausecanbeusedtonameasubquery.Whichofthefollowingis
alsotrue?(Choosetwo.)
A.ThenameofthesubquerycanbeusedintheSELECTstatement
followingtheWITHclause.
B.Thenameofthesubquerycanbejoinedtoothertablesinthe
SELECTstatementfollowingtheWITHclause.
C.ThenameofthesubqueryisstoredinthedatabasebytheWITH
statementandcanbereferencedbyotherSQLstatementsinlater
sessions.
D.Thenameofthesubquerycanbeinvokedfromwithinthe
subquerythatisnamed.
15.Reviewtheillustrationfromquestions8.Whichofthefollowingstatements,
whenexecuted,willresultinanerror?
WriteSingle-RowandMultiple-RowSubqueries
16.Whichofthefollowingcomparisonoperatorscanbeusedwithamultiple-
rowsubquery?(Choosetwo.)
A.=
B.>=ALL
C.LIKE
D.IN
17.ReviewthePORTSandSHIPStables:
Next,reviewthefollowingSQLcode:
YouknowthattherearefiverowsintheSHIPStablewithalength
greaterthan900.WhatwillresultfromanattempttoexecutethisSQL
statement?
A.Anexecutionerrorwillresultbecausethesubquerywillreturn
morethanonerowandtheparentqueryisexpectingonlyonerow
fromthesubquery.
B.AsyntaxerrorwillresultbecausePORT_IDand
HOME_PORT_IDinline3havedifferentcolumnnames.
C.Thestatementwillexecuteandproduceoutputasintended.
D.Noneoftheabove.
18.Whenisaqueryconsideredamultirowsubquery?(Choosethebestanswer.)
A.Ifitreturnsmultiplerowsatthetimeofexecution
B.Ifitmayormaynotreturnmultiplerows,asdeterminedbyits
WHEREclause
C.Ifitreturnsnumericdata,regardlessofthenumberofrowsof
dataitreturns
D.Alloftheabove
19.Whichofthefollowingisatruestatement?
A.IfaSELECTincludesaGROUPBYclause,thenanysubquery
usedwithintheSELECTmustalsohaveaGROUPBYclause.
B.Ifaqueryreturnsmultiplerows,itmaynotbeusedasa
subqueryforaSELECTstatementthatusesaGROUPBYclause.
C.ASELECTstatementwithaGROUPBYmayuseasubqueryto
returnavaluetotheoutermostWHEREclause.
D.TheonlyformofsubquerypermittedwithaGROUPBYclause
isacorrelatedsubquery.
20.ReviewthePORTSandSHIPStablesshowninquestion7.Thenreviewthe
followingSQLcode:
Whichofthefollowingistrueofthisstatement?
A.Thestatementwillfailwithasyntaxerrorbecauseofline3.
B.Thestatementwillfailwithasyntaxerrorbecauseofline5.
C.Whetherthestatementfailsdependsonhowmanyrowsare
returnedbythesubqueryinlines3through5.
D.Noneoftheabove.
SELFTESTANSWERS
DefineSubqueries
1. A.Ascalarsubqueryalwaysreturnsasinglevalue,whichistosayit
returnsonlyonecolumn,onerow.
B,C,andDareincorrect.Acorrelatedsubquerymayormaynotreturn
multiplerows.Multiple-columnsubqueriesmayormaynotreturnmultiple
rows.
DescribetheTypesofProblemsSubqueriesCanSolve
2. BandC.Theidealquerywouldfirstobtaintheaveragesalesforthe
prioryearandthenusetheresultstocomparetothesubsequentyear’s
sales.Youuseasubquerytocreateaview.
AandDareincorrect.Theminimumsalesateverydivisioncanbe
determinedusingaquerywithaGROUPBY.Asubqueryisnotrequiredor
particularlynecessaryhere,basedontheinformationprovided.Youdonot
useasubquerytocreateasequence.
DescribetheTypesofSubqueries
3. BandD.Asingle-rowsubquerymayconsistofmultiplecolumnsin
itssinglerow.Anditalsomaybecorrelated.
AandCareincorrect.Asingle-rowsubquerycannot,bydefinition,also
beamultiple-rowsubquery.Ascalarsubquerybydefinitioncanbeonly
onecolumninonerow,soitcannotbeamultiple-columnsubquery.
4. B.Acorrelatedsubqueryisthebestanswer.Thenameindicatesthat
thesubqueryiscorrelatedtotheparentquery.Technically,scalarand
multiple-columnsubqueriesmayalsobecorrelated,butweaskedforthe
bestanswer,andclearlythatiscorrelatedsubquery.
A,C,andDareincorrect.Technically,ascalarsubquerymayalsobea
correlatedsubquery,whichiswhythequestionaskedyoutopickthe“best
answer”—thetermcorrelatedrefersspecificallytotheconceptofreferring
totheparentquery,whereasascalarsubqueryrefersspecificallytoa
subquery’sreturnvalueasbeingasinglevalue.Thesameistruefora
multiple-columnsubquery—thatmayalsobeacorrelatedsubquery,butthe
term“multiple-column”isintendedtoemphasizethefactthatitreturns
multiplecolumns’worthofresultsatonce.Finally,areferentialsubquery
isn’tanything;wejustmadethatup.
5. A,C,andD.AsubquerycouldbeusedwithintheSELECT
statementofanINSERTstatement,includingtheSELECTstatement’s
WHEREclause,andwithinaninlineview.
Bisincorrect.AGRANTstatementisaDCLstatementandcannothave
asubquery.
6. A.Aninlineviewiscalledaviewsinceitisaquerythatistreated
likeatable.Thatbeingsaid,itssyntaxisthatofthesubquerysyntax.
Bisincorrectbecausethestatementistrue.
QueryDataUsingCorrelatedSubqueries
7. AandC.Aisaclassiccorrelatedsubquery,connectingthesubquery
totheparentbywayoftheW1.SHIP_IDvalue.Calsoworkswiththe<=
ALLcomparisoncondition.
BandDareincorrect.Bismissingthejoininthesubquerythatconnects
thesubquerywiththeparentquery.Dcomparestheparentquery’sWHERE
clausevaluetothesubquerywithaless-thansign,whichwon’twork—the
subqueryisalreadyselectingtheminimumvaluefromthesubquery,sothe
parentquerycan’tfindanythingthatwillbelessthantheminimum.
8. D.Thestatementissyntacticallyfine.TheSELECTincludestwo
correlatedsubqueries.Thefirst,inlines1through3,isanexpressioninthe
SELECTstatement’sselectlist.Thissubqueryiscorrelatedbywayofthe
referenceattheendofline3.Thesecondcorrelatedsubqueryisinlines5
through7,anditobtainstheminimumcapacityvalueforshipsbelongingto
eachport.
A,B,andCareincorrect.Thestatementwillnotfail,notwithasyntax
problemorwithanexecutionproblem.Thedataitreturnsisexactlyas
requested.
9. B.Acorrelatedsubquerycannotbeexecutedasastand-alonequery.
A,C,andDareincorrect.Acorrelatedsubquerycanbeusedina
SELECT,UPDATE,orDELETE.Acorrelatedsubquerydoesnothaveto
useatablealiaswhenreferencingcolumnsunlessthereisariskof
ambiguity.
UpdateandDeleteRowsUsingCorrelatedSubqueries
10. A,B,andC.Acorrelatedsubquerycanbeusedinanyoftheseanswers.
Disincorrect.AcorrelatedsubquerycannotbeusedintheFROMclause
ofaDELETEstatement.Note,however,thatthequestionisasking
specificallyaboutcorrelatedsubqueries.Whileyoucannothavea
correlatedsubqueryintheFROMclauseoftheDELETE,we’lllatersee
thatan“inlineview”canbeusedthere,andan“inlineview”isessentiallya
subquery—butnotacorrelatedsubquery.
11. A.TheCAPACITYwillincreaseonceforeachrowprocessedbythe
UPDATEifthatrowisfoundinthesubquery.
B,C,andDareincorrect.
UsetheEXISTSandNOTEXISTSOperators
12. D.Thesemijoinisthecorrectanswer.
A,B,andCareincorrect.Thereisnosuchthingasademijoin.A
multiple-columnsubqueryrequiresseveralcolumnsonbothsidesofthe
comparisoncondition.Across-joinisatablejoinwithnojoincriteria.
13. A.Itwillfailbecauseofasyntaxerroronline2—thefirstreferenceto
PORT_IDshouldberemoved.EXISTSdoesnotcomparethesubqueryto
anything.Inotherwords,line2shouldbeWHERENOTEXISTS
(SELECT*withoutthefirstPORT_IDreference.Otherthanthat,
everythingelseaboutthequeryisfine.
B,C,andDareincorrect.Line4hasnosyntaxerrors.Nordoesthe
subquerycontainanyexecutionerrors.ButneitherwilltheSQLexecute,
forthereasonswedescribedfortherightanswer.
UsetheWITHClause
14. AandB.ThenamecanbeusedintheSELECTfollowingtheWITH
clause.
CandDareincorrect.Thenameisnotstoredinthedatabasebythe
WITHstatement.ItexistsonlyfortheWITHclauseitselfandisnot
recognizedoutsideoftheWITHclause.TheoneplacewithintheWITH
clausethatdoesnotrecognizethesubquerynameiswithinthenamed
subqueryitself.
15. AandD.TheWITHclausemustnamethequeryinorderfortheSELECT
tobeabletoreferencethequerycontainedwithintheparentheses.The
syntaxofWITHdoesnotusetheSELECTprecedingtheWITH.
BandCareincorrect.Eachissyntacticallycorrect.Bothnamethequery
intheWITHclause.
WriteSingle-RowandMultiple-RowSubqueries
16. BandD.The>=ALLsyntaxis“greaterthanorequaltoall”thevalues
returnedbythesubquery,whichisidealforamultiple-rowquery.TheIN
comparisonoperatorisalsouseful.
AandCareincorrect.Theequalsignisrestrictedtoonlysingle-row
subqueries.TheLIKEoperatorisalsolimitedtosingle-rowsubqueries.
17. A.Thequerywillproduceanexecutionerrorbecausetheparentqueryis
expectingasingle-rowanswerfromthesubquery.Youknowthisbecause
ofthecomparisonoperatorinline3:thegreater-thansignisasingle-row
comparisonoperator.Thebetterchoiceheremightbe>ANYor>ALL,
dependingonthesituation.
B,C,andDareincorrect.ThereisnothingwrongwithPORT_IDand
HOME_PORT_IDhavingdifferentcolumnnames.Aslongastheirdata
typesmatch,alliswell,andyouknowtheirdatatypesmatchaccordingto
theillustrationprovided.
18. A.Itreturnsmultiplerowsatthetimeofexecution.
B,C,andDareincorrect.TheWHEREclauseiscertainlyimportant,but
thequery’sparticularformisnotimportant.Whatisimportantiswhether
thequeryreturnsmultiplerowsatthetimeitexecutes.Thepresenceof
numericdataisnotimportant,noristhenumberofcolumns.
19. C.AsubquerycanbeusedtoreturnavaluetotheWHEREclauseofa
SELECTstatement,regardlessofwhetheraGROUPBYisinvolved.The
issueofwhetherthesubqueryreturnsonerowormultiplesmustbe
handledcorrectly,aswe’vediscussedelsewhere.Butprovideditisdone
correctly,asubquerycanbeusedinthiscontext.
A,B,andDareincorrect.ThereisnoissuewithregardtoGROUPBY
andtheuseofsubqueries.ASELECTwithaGROUPBYisallowedthe
optionoftheuseofsubqueries,eithersingle-rowormultiple-row(orboth)
providedthattherulesofthesingle-rowand/ormultiple-rowsubqueryare
handledcorrectly.AGROUPBYispermittedtouseanyformofsubquery;
itisnotrestrictedtothecorrelatedsubquery.
20. A.Thestatementwillfailwithasyntaxerrorbecausetherearetwo
columnsinthesubquery,butonlyonevalueisontheleftsideofthe
comparisonoperatorIN.
B,C,andDareincorrect.Thesyntaxonline5isacceptable.Theuseof
INinline3ensuresthatthesubquerywillbepermittedtoreturneitherone
ormultiplerows,andwithallelsebeingacceptable,thestatementwould
otherwiseexecutesuccessfully.
10
ManagingSchemaObjects
I
CERTIFICATIONOBJECTIVES
10.01DescribeHowSchemaObjectsWork
10.02CreateSimpleandComplexViewswithVisible/InvisibleColumns
10.03Create,Maintain,andUseSequences
10.04CreateandMaintainIndexesIncludingInvisibleIndexesandMultiple
IndexesontheSameColumns
10.05PerformFlashbackOperations
✓Two-MinuteDrill
Q&ASelfTest
nadditiontotables,thereareotherimportantschemaobjects.Thischapter
looksatseveral,includingviews,sequences,andindexes.Thischapteralso
addressesadditionalfeatures:FlashbackTable,FlashbackDrop,andFlashback
Query.
CERTIFICATIONOBJECTIVE10.01
DescribeHowSchemaObjectsWork
Asyou’veseen,thereareseveralobjectsinthedatabase.Manyare“schema”
objects,meaningtheyare“owned”byauserandexistinacollectionwithina
useraccount.Schemaobjectsincludetables,views,indexes,andsequences.In
thissection,we’lllookatsomeofthefunctionalityofdatabaseobjectsandhow
theyworkwitheachother.
Tables
Allthedatainadatabaseisstoredintables.Whenyoucreateanewtable,the
informationaboutthattable,suchasitsnameandcolumnsandthedatatypesof
thosecolumns,isallstoredinasetofsystem-definedtablesthatarecollectively
knownasthedatadictionaryandthataremanagedautomaticallybytheOracle
system.(WeexaminethedatadictionaryinChapter12.)So,evendataabout
yourdata—inotherwords,metadata—isstoredintables.
Atable’scolumnsaregenerallystoredintheorderinwhichtheyarecreated,
butthisisn’tnecessarilytrueatalltimes.Also,ifyoualteratableviaALTER
andaddacolumntoitsstructure,thenewcolumnwillbeaddedattheendofthe
listofcolumnsinthetable’sstructure.
Constraints
Aconstraintisaruleonatablethatrestrictsthevaluesthatcanbeaddedtoa
table'scolumns.Notethatitisnotadatabaseobject,butitislistedinthedata
dictionaryandcanbenamedwiththesamenamingrulesofanobject.You’ve
alreadylookedatthedifferenttypesofconstraints:NOTNULL,UNIQUE,
PRIMARYKEY,FOREIGNKEY,andCHECK.
IfaSQLstatementattemptstoadd,modify,ordeletearowto/fromatable
andinsodoingviolatesaconstraint,theentireSQLstatementwillfailwithan
executionerror.
Views
Aviewactslikeatable.Ithasaname.Youcandescribetheviewwith
DESCRIBEinthesamewayyouwoulddescribeatable.YoucanrunaSELECT
statementonaviewjustasyouwouldselectfromatable.Dependingonthekind
ofviewyouareworkingwith,youmightevenbeabletoexecuteINSERT,
UPDATE,and/orDELETEstatementsonaviewtomanipulatethecontentsof
theunderlyingtables.
Butaviewisnotatable,anditstoresnodata.Aviewisnothingmorethana
SELECTstatementthatissavedinthedatabasewithanameassignedtoit.The
columnstructurethattheviewtakesonisformedbytheSELECTstatement’s
selectlist.Thedatatypesforthosecolumnsarepickedupautomaticallyfromthe
underlyingtableortheexpressionsusedintheSELECTstatementtocreatethe
view.
Indexes
Anindexperformsthesamejobthatatypicalindexinabookperforms.For
example,OracleCorporation’sSQLLanguageReferenceManualforOracle
12c’sR1releaseismorethan1,908pageslong.Whatifyouwerelookingfor
informationontheDISTINCTclauseoftheSELECTstatement?Youhaveafew
waystofindsuchinformationinthebook.Onewayistositdownandstart
readingthebookatthefirstpageandkeepreadinguntilyoufindthedatayou’re
lookingfor.Amuchmoreefficientwayistofliptothebackofthebookandfind
theindex,whichcontainsasummaryofimportanttopicsinalphabeticalorder.
WithinafewsecondsyoucanlookupthewordDISTINCT,notethepage
numberonwhichit’smentioned,andflipstraighttoit.That’samuchbetter
approach.
TheSQLINDEXobjectperformsinmuchthesameway.Whenyoucreatean
INDEXobject,youareidentifyingoneormorecolumnsinatablethatyou
believewillbefrequentlyusedtolookupdata.Thenyoucreatetheindexbased
onthatcolumn—orsetofcolumns—andOracleliterallybuildsaseparateobject
thattakesauniquelistofallthedatacurrentlyinthatcolumn,sortsit
appropriatelyaccordingtodatatype,andthenstoresinternaladdressing
informationthattiestheindextothesourcetableandtherowscontainedwithin.
Theresultisthatanyfuturequeriesonthetablethathappentoreferenceany
indexeddatawillcausethefollowingtooccurautomatically:
Performananalysistodeterminewhetherthequerywillbenefitby
usingtheindex
Ifyes,thenredirectthefocustemporarilytotheindex,searchthe
indexforanyofthedesireddataidentifiedbythequery,andobtaindirect
locationsoftheappropriaterows
Thedifferenceinperformanceispotentiallysignificant.Themoredatathatis
storedinatable,themorebeneficialanindexmaybe.
Sequences
Asequenceisacounter,meaningthatitissuesnumbersinaparticularseries,
alwayskeepingtrackofwhateverthenextnumbershouldbe.Ifyouever
watchedtheclassictelevisionsitcomSeinfeld,youmightrecalltheepisode
whereJerryandElainegotoabakerytopickupachocolatebobka,butdon’t
knowtotakeanumberfromthedispenserwhentheyfirstarriveandlosetheir
placeinlineasthebakerygetscrowded.Thatdispenser—whichissuespaper
ticketsidentifyingtheholderasbeing,say,“number42”inline—servesthe
samepurposeofaSEQUENCEobjectinthedatabase.
TheprimarypurposeofSEQUENCEistosupporttheprocessofaddingrows
toaparticulartableandprovidingthenextuniquevalueforaPRIMARYKEY
forthetable.That’sit.There’snothinginherentinaSEQUENCEobjectthatties
ittoaparticulartable.There’snothingautomaticintheSEQUENCEobjectthat
necessarilysupportsaparticulartable.It’suptothesoftwaredevelopertoknow
howtousetheSEQUENCEobjectcorrectly.Wewillexplorethisindetaillater
whenweaddressthesyntaxforworkingwithaSEQUENCE.
CERTIFICATIONOBJECTIVE10.02
CreateSimpleandComplexViewswith
Visible/InvisibleColumns
AviewisaSELECTstatementwithaname,storedinthedatabase,and
accessibleasthoughitwereatable.EarlieryousawtheWITHclausethatcan
assignanametoaquerywithinasingleSELECTstatement.Theviewobject
doesthesamethinginamorepermanentmanner,meaningthattheviewobject
residesinthedatabasealongsidetablesandotherdatabaseobjects.
Onceyou’vecreatedaview,youcanrefertoitinSELECTstatementsas
thoughitwereatable.NothingabouttheSELECTisdifferent—anyonelooking
atagivenSELECTstatementwillnotbeabletodeterminefromtheSELECT
statementaloneiftheFROMclausespecifiesatableoraview.
Viewsareusefulforavarietyofreasons.Onebenefitissecurity.For
example,consideratypicalscenariowhereyouhavealargetablethatcontainsa
combinationofsomesensitiveinformationalongwithsomeinformationthatis
ofageneralinterest.Youmightcreateaviewthatqueriesthegeneral-interest
columnsofthelargetableandthengrantprivilegesonthatviewtothegeneral
populationofusers.Thoseusersmaynowquerytheviewandgetdirectaccess
tothegeneralinformationwithouthavingaccesstomoresensitivedatathat
existsintheunderlyingtable.Viewsareagreatwaytomaskcertaindatawhile
givingaccesstootherdatainthesametable.
Anotherbenefittoviewsistheirabilitytomakeacomplexqueryeasierto
workwith.Forexample,youmightcreateaviewthatisbuiltonacomplexjoin
sothatthecomplexityisbuiltintotheview.Theresultisaviewobjectthat
appearstobeasingletable,whichyoumaynowqueryasthoughitwereatable.
Youcanevenjointheviewwithothertablesandotherviews.Inthissituation,a
viewcanbeusedtosimplifythecomplexityofacommonlyusedjoin.
Inthenextsection,we’llcreateaviewobject.
CreatingViews
Let’slookatanexample.First,reviewFigure10-1.We’llstartwithjustthe
EMPLOYEEStable—noticethatitincludescolumnsforemployeeID,name,
SocialSecuritynumber,dateofbirth,andprimaryphonenumber.
FIGURE10-1 DiagramsfortheEMPLOYEESandPAY_HISTORYtables
So,here’saproblem:whatifyouwantedtogiveaccesstothistablesothat
otherpeopleintheorganizationcangetthephonenumbersofemployees?Think
aboutthatsensitiveinformation,includingSocialSecuritynumbers,andyou
mighthavesecondthoughtsabouthavinganybodyquerytheEMPLOYEES
table.
Onesolutiontothispredicamentistouseaview.Let’screateaviewforthe
EMPLOYEEStable.
IfweexecutethisstatementinSQL,we’llgetthefollowingmessage:
(Note:ThisstatementisthemessagedisplayedbyOracleSQLDeveloper.
SQL*Pluswilldisplaya“Viewcreated”message.)
Nowthatwehavethisview,wecanworkwithitjustasifitwereatable.For
example,wecanuseDESCRIBEtoexplainit.
Additionally,wecanselectfromitwithSELECTasifitwereatable.
Theresultswilldisplayjustlikeanytable.Anyoneusing
VW_EMPLOYEES,runningSELECTstatementsonit,describingitsstructure,
anddoinganythingtheymightwantusingSQLcommandsmaynoteverrealize
it’snotatableatall.Thefactthatit’saviewisn’tnecessarilyobvious.Wechose
tonameitwithaVW_prefixforourownpurposes,butwecouldhavegivenit
anynamewewanted.
Thesyntaxrulesforcreatingavieware
ThekeywordsCREATEVIEW
TheoptionalkeywordsORREPLACE
Anamefortheview,specifiedaccordingtotherulesfornaming
databaseobjects
ThekeywordAS
Finally,avalidSELECTstatement,withafewrestrictions
OneoftherequirementsoftheCREATEVIEWstatementisthis:theresulting
VIEWmusthavevalidcolumnnames.ThismeansthatifyouchooseaSELECT
statementthatincorporatesanycomplexexpressionswithintheselectlist,each
expressionmustbeassignedacolumnalias,orcolumnnamesmustbeassigned
bytheCREATEVIEWstatementitself.Forexample,let’susetheOR
REPLACEoptiontocreateanewversionofourVW_EMPLOYEESview(line
numbersadded).
Whatwentwrong?Theproblemisonline3,wherewespecifytheexpression
thatformsthesecondcolumninourSELECTexpressionlist.Noticethatit
concatenatesthecolumnLAST_NAMEwithacomma,aspace,andthenthe
FIRST_NAMEcolumn.Here’stheproblem:whatisthenameofthiscolumn?
Thereisn’toneassigned,buttheCREATEVIEWstatementrequiresone.
Therefore,theattemptedCREATEVIEWstatementfails.
Tocorrecttheproblem,wecanspecifycolumnnamesintheCREATEVIEW
statement,likethis:
Nowwe’llhaveaviewconsistingofthreecolumnswiththenamesID,
NAME,andPHONE.Alternatively,wecouldhaveusedacolumnaliasinthe
SELECT
Youcanusecomplexqueriestocreateviews.Forexample,let’sgobackto
Figure10-1andbuildaquerythatjoinsdatafromthePAY_HISTORYtableinto
ourview,asfollows:
ThisstatementusesaSELECTstatementwithajoin,aGROUPBY,a
WHEREclause,andaggregatefunctions;itcreatesaVIEWobjectoutofallof
it.
VIEWobjectscanbebasedonSELECTstatementswithsubqueries,
functions,andmore.
Theviewcreatedwiththisstatementwillconsistonlyoftherowsinthe
SHIP_CABINStablewheretheSHIP_IDvalueisequalto1.The
SHIP_CABINStablemaycontainmorerowsthanthosewithaSHIP_IDof1,
butthoserowswon’tbefoundbywayofanyqueryontheSHIP_ONE_CABINS
view.
Also,takenoteagainoftheoptionalkeywordsORREPLACEinthe
CREATEVIEWstatement.Thesewordsdon’tworkwithCREATETABLE,but
theydowithCREATEVIEW.Becarefulwiththem—whenincluded,theywill
giveyounowarningifyouareorarenotoverwritingsomeexistingview.They
willsimplyoverwritewhatmayhavegonebeforeandreplaceitwiththenew
view.It’saconvenientoptionthatispowerful,sobecarefulwithit.Youhave
beenwarned.
ViewsandConstraints
Youcancreateconstraintsonaviewinthesamewayyoucancreatethemona
table.However,Oracledoesn’tenforcethemwithoutspecialconfigurationthat’s
availableprimarilytosupportcertaindatawarehousingrequirements.This
subjectisworthnotingbutnotspecificallyaddressedontheexam.
UpdatableViews
You’veseenhowwecanSELECTfromaview.ButcanwealsouseINSERT
andUPDATEandDELETE?
Theansweris,itdepends.Iftheviewcontainsenoughinformationtosatisfy
alltheconstraintsinanyunderlyingtables,thenyes,youcanuseINSERT,
UPDATE,orDELETEstatementsontheview.Otherwise,no.
Dependingonthenatureoftheconstraints,itmaybepossibletousesome
DataManipulationLanguage(DML)statementsbutnotothersontheview.For
example,iftheviewfailstoincludeanycolumnsfromtheunderlyingtablethat
haveNOTNULLconstraints,youwillnotbeabletoexecuteanINSERT
statement,butyoumaybeabletoissueanUPDATEorDELETEstatement.
Forexample,let’srevisittheEMPLOYEEStableinFigure10-1.Let’smake
anewviewonthattable,likethis:
Thisviewcontainsjustenoughinformationtoprintanemployeephonebook,
includingtheirnamesandphonenumbers.Oncewe’vecreatedthisview,wecan
selectfromit,likethis:
Fantastic.Butwait—wejusthiredsomeonenew.HernameisSonia
Sotogovernor,andwewanttoaddhernametothisview.Let’stryit.
Whoops.What’swronghere?Seetheerrormessage?Rememberthatour
underlyingtableistheEMPLOYEEStable,anditcontainsaPRIMARYKEYof
EMPLOYEE_ID,asshowninFigure10-1.Butourviewdoesn’tincludethat
column.Asaresult,thereisnowayforustoinsertavaluethroughthe
EMP_PHONE_BOOKviewsothatitwillprovideavaluefortherequired
columnofEMPLOYEE_ID.WecannotexecuteanINSERTstatementonthis
view.
However,wecanexecuteanUPDATEstatementontheview.Here’san
example:
Thisstatementworksperfectlyfineontheview.Thereasonisthatthe
changeswe’remakingwiththisUPDATEstatementdonotviolateany
underlyingconstraints.
Ifthereareanyconstraintsontheunderlyingtablethatyoucannotsatisfy
whenattemptingtoissueanINSERT,UPDATE,orDELETEonaview,thenthe
statementwon’twork.TheVIEWobjectmustprovideaccesstoanyofthe
underlyingtable’scolumnsinsuchawaythattheconstraintscanbehonoredto
satisfythoseconstraintsandexecutethestatementsuccessfully.Otherwise,the
INSERT,UPDATE,orDELETEstatementwillfail.
Inaddition,aviewthatisbasedonaggregatedrowswillnotbeupdatable.
YouwillbepreventedfromusingINSERT,UPDATE,orDELETEifyou
createaviewbasedonaSELECTstatementthatincludesanyofthefollowing:
Omissionofanyrequiredcolumns(NOTNULL)inthatunderlying
table
GROUPBYoranyotheraggregation,suchassetoperators
DISTINCT
AFROMclausethatreferencesmorethanonetable—thatis,
subqueriesintheSELECTormostjoins
Regardingthatlastitem,itistechnicallypossibletoexecuteDMLchangeson
joinswhereallupdatablecolumnsbelongtoakey-preservedtable.Thedetails
gobeyondthescopeofthisbook.Forthemostpart,youwillnotbeabletoissue
DMLchangestoaVIEWobjectbasedonajoin.
Aswe’vealreadyseen,aview’sSELECTstatementmayincludeexpressions
aspartofthecolumnsinitsformation,asfollows:
NotethattheprecedingqueryconcatenatestheFIRST_NAMEand
LAST_NAMEcolumnsintooneexpression.Asaresult,theindividualcolumns
cannotbemodifiedwithanINSERTorUPDATEstatement—thereisnowayto
singularlyrefertotheindividualcolumnsunlesstheyareaddedasindividual
itemsintheselectlist.However,EMPLOYEE_ID,therequiredcolumn,is
includedasanindividualcolumn,sothiswouldbeasatisfactorystatement:
ThatstatementwillsuccessfullyexecuteonourEMP_PHONE_BOOKview
andaddanewrowtotheunderlyingtable,assumingtheprimarykeyvaluefor
EMPLOYEE_IDisacceptedasanewuniqueentry.Butwe’renotabletoinsert
arowthroughtheviewusinganINSERTstatementthatreferencesthe
EMP_NAMEcolumnaliasoritscomponentcolumnsFIRST_NAMEand
LAST_NAME.Wesimplyhavetoomitanyreferencestothosecolumnsinour
DMLstatementsforourDMLtoexecutesuccessfully.TheINSERTINTO
EMP_PHONE_BOOKstatementwillcauseanewrowtobeaddedtothe
EMPLOYEEStable.ThatnewrowwillconsistofavalueforEMPLOYEE_ID
(102)andavalueforPRIMARY_PHONE('800-555-1212');allothercolumnsin
thenewrowwillshowasNULL—thatis,havenovalues.(Remember,NULL
representstheabsenceofinformation.)
Wemaydeletearowinthisview.Here’sanexample:
ThisstatementwillsuccessfullydeletetheentirerowforanEMPLOYEE_ID
of102.IfweissuedasimilarDELETEstatementforanyotherexistingvaluewe
canaccess,suchasPRIMARY_PHONEorEMPLOYEE_ID,andtherowwere
found,thentherow—theentirerowoftheunderlyingtable—wouldbedeleted.
Thatincludesdataincolumnswecan’tevenseewiththeview.Thewholerow
willbedeleted.
WithregardtothegeneralquestionofusingINSERT,UPDATE,and/or
DELETEonanygivenview,thegeneralanswerissimple:iftheviewprovides
row-level(notaggregated)accesstoone—andonlyone—tableandincludesthe
abilitytoaccesstherequiredcolumnsinthattable,thenyoucanuseINSERT,
UPDATE,and/orDELETEontheviewtoeffectchangestotheunderlyingtable,
inaccordancewiththerestrictionslistedearlier.Otherwise,youmaynotbeable
tosuccessfullyexecuteachangetotheview’sdata.
NotethattheINSTEADOFtriggerinPL/SQLcanbeusedtocausea
nonupdatableviewtobehaveasthoughitwereupdatable.ButPL/SQL
featuresarenotaddressedontheexam.
InlineViews
AninlineviewisasubquerythatiscontainedwithinalargerSELECTstatement
insuchawaythatitreplacestheFROMclauseofaSQLstatement.
Here’sanexample:
Inthisexample,theinlineviewisincludedintheparenthesesattheendof
line2.
Thereisnolimittothenumberofinlineviewsyoucannestwithininline
views.
This“unlimitednesting”isdifferentfromthelimitfortypicalsubqueries,
wherethelimitis255nestedsubqueries.
Inlineviewscanbecombinedwithvariouscomplexqueries,suchasthose
thatuseJOINandGROUPBYclausesandmore.Here’sanexample:
ThisstatementisasingleSELECTthatpullsdatafromtwoinlineviews,one
onlines2through4andthesecondonlines6through8.
InlineviewscanbeanyvalidSELECTstatement,placedintoaSQL
statementwheretheFROMclausewouldnormallygo.
Onegreatusageofaninlineviewistoaddressanissueinvolvingthe
pseudocolumnROWNUM.ROWNUMisarownumberautomaticallyassigned
toeachrowresultofaquery.ThechallengewithROWNUMisthatitis
assignedbeforetheORDERBYclauseisprocessed.Asaresult,youcannotsort
rowsandthenuseROWNUMasawayofdisplaying,forexample,theline
numberforeachrowofoutput.TheresultswillbemixedupsinceROWNUMis
determinedbeforeORDERBYisprocessed.ButyoucanmovetheORDERBY
clauseintoaninlineviewandthenusetheROWNUMpseudocolumnonthe
outerquerytodisplayrownumberscorrectly.Here’sanexample:
Inthisexample,weuseORDERBYintheinlineviewtosortourrowsby
INVOICE_DATE,andthenweuseROWNUMintheouterquerytolimitour
outputtojustthefirstthreerowsofdata.WealsoincludeROWNUMinour
selectlistsothatitappearsintheoutput.Withouttheinlineview,oddsarethat
ourROWNUMvalueswouldbeinanapparentlyrandomorderinsteadof
sequential.
So,whywouldyouwanttouseaninlineview?Therearemanyreasons.As
we’vealreadydemonstrated,inlineviewscanbeusedtocreatecomplexjoins
builtonaggregatedqueries.Anotherbenefithastodowiththenatureof
dynamicSQLasit’susedwiththird-generationlanguages.Manypopularweb
sitesarebuiltonwebpagesthataredynamicallyformedfromacombination
ofJava,PL/SQL,orPHPcodethatpullsdatafromthedatabaseandmerges
theoutputwiththelanguagesusedtoformwebpages.Afullexampleofsuch
ascenarioisbeyondthescopeofthisbook,butit’sworthnotingthatsuch
systemsrelyheavilyonroutinesthatcreateSQLcodedynamically,during
execution,inresponsetoqueriesfromendusers.Suchdynamicscenarioscan
benefitgreatlyfromtheabilityto,forexample,createastandardouterquery
inwhichtheinlineviewcanbesubstitutedbydynamiccode.Anendusermay
performasearchthatmightdrawdatafromanynumberofvarioussources
yetpresenttheoutputthroughafixedseriesofdataelements.Theinlineview
cansupportsuchasituation.
ALTERVIEW
TheALTERVIEWstatementisusedtoaccomplishanyofthefollowingtasks:
Create,modify,ordropconstraintsonaview
Recompileaninvalidview
Thesubjectofconstraintsonaviewisnotsomethingthatiscoveredonthe
exam.Oracledoesnotenforceviewconstraintswithoutspecialconfiguration
(seeDISABLENOVALIDATEinOracle’sdocumentation).
Recompilingaviewisastepyoumaywanttotakeifyou’vecreatedaview
andthenlaterperformedsomesortofmodificationontheunderlyingtableor
tablesuponwhichtheviewiscreated.Dependingonthechangeyoumaketothe
view’ssourcetable,theviewmayberenderedinvalidasaresult.Oncerendered
invalid,asubsequentattempttousetheviewwillresultinoneoftwooutcomes,
dependingonthenatureofthechangethatwasperformedtotheunderlying
table:
Outcome1Ifthechangeresultsinasituationwheretheoriginal
CREATEVIEWstatementwouldhavesuccessfulexecutedanyway,then
asubsequentstatementusingthatviewwillcausetheviewtobe
recompiledsuccessfully,andthestatementwillsuccessfullyexecuteas
thoughnoproblemeverexisted—becauseitdidn’t.
Outcome2Ifthechangeresultsinasituationwheretheoriginal
CREATEVIEWstatementwouldhavenotsuccessfullyexecuted,then
subsequentattemptstousetheviewwillfail.Anexampleofsucha
changewouldbetoalteranunderlyingtabletodropacolumnthatispart
ofaview’scolumnlist.
InChapter12we’llseehowyoucandeterminewhetheraviewisinvalidor
notbyqueryingthedatadictionary.
Analternativetoqueryingthedatadictionary,whichisnotalwaysavailable
toanapplication,istoissueanALTERVIEW…COMPILEstatement.This
statementcanbeusedtorecompileaninvalidview.Iftheviewcannotbe
recompiled,theALTERVIEW…COMPILEstatementwillsuccessfully
executewithamessagedeclaring“Warning:Viewalteredwithcompilation
errors.”Thedatadictionarywillshowtheviewasremaininginvalid.
Hereisanexampleofastatementthatrecompilesaview:
Youcannotchangeaview’sSELECTstatementwiththeALTER
VIEWstatement.Instead,youmustdropandre-createtheview.
Oncecompiledsuccessfully,theviewisbackinworkingcondition.Ifitdoes
notcompile,youknowthatthechangetotheunderlyingtablemayhave
fundamentallychangedthenatureoftheview’sstructure.Forexample,ifaview
queriesanamedcolumnfromanunderlyingtableandthatcolumnisrenamed
(ordropped,asIpointedoutearlier),thentherecompilationwillnotwork,and
youmayneedtore-createtheviewandreassessyourcode.
Inaddition,ifyousimplyattempttouseaninvalidview,itwillautomatically
compile.Ifitcanbecompiled,theattempttousetheviewwillbesuccessful,
withallelsebeingequal.However,ifitcannotberecompiled,thencontraryto
theALTERVIEW…COMPILEstatement,whichwouldexecutebutissuea
warning,anyattempttouseaninvalidviewthatcannotbeautomatically
recompiledwillresultinanerror.
Visible/InvisibleColumns
Viewsarebasedontables,andinOracle12c,atablecanhavecolumnsthatare
eithervisibleorinvisible.Howdoesthatimpactviews?Let’sstartbylookingat
theuseofinvisiblecolumnsintables.
InvisibleColumnsandTables
Thecolumnsinatablearevisiblebydefault.NewwithOracle12c,youcan
specifythatoneormorecolumnsbeinvisible.Here’sanexample:
NotethecolumnCONSTRUCTION_COST,whichisspecifiedas
INVISIBLE.Theothercolumnsarevisiblebydefault;onlythecolumn
CONSTRUCTION_COSTisinvisible.
So,whatdoesthismean?First,let’slookatdisplayingthetable’sstructure
withDESC.YoucandescribethetablewithDESCSHIP_ADMIN,andyouwill
seethetypicaltabledescription,butanycolumnspecifiedasINVISIBLEwillbe
omittedfromdisplay.
NotethatwithinOracle'sSQL*Plustoolspecifically,youcanusethe"SET
COLINVISIBLEON"statementtotemporarilyconfiguretheDESCstatement
toreturninvisiblecolumnsduringthatparticularSQL*Plussession.
YoucanuseINSERTwithatablethathasinvisiblecolumns,butyouhaveto
becareful.YoucanuseINSERTtoinsertvaluesintoaninvisiblecolumn,but
onlyifyouspecifythecolumnnames.IfyouomitthemintheINSERT,you’ll
getanerrormessage.Foranexample,seeFigure10-2.
FIGURE10-2 TheINSERTstatementandcolumnsthatarespecifiedas
INVISIBLE
NotethefirstINSERTstatementinthefigure,whichusestheshorthand
INSERTsyntaxthatomitsatable’scolumnsbeforetheVALUESclause;you’ll
getanerrormessageifaninvisiblecolumnispresentinthetable.Thisshorthand
syntaxisanimpliedreferencetoeverycolumninthetable.Inthiscase,the
SHIP_ADMINtablehasthreecolumns,butonlytwoarevisible.Thefinal
portionofthisfirstINSERTprovidesvaluesforthreecolumns,andthetablehas
threecolumns—butonlytwoarevisible.WiththeshorthandINSERTsyntax,
thisstatementfails.HadweusedashorthandversionofINSERTthatprovided
onlytwovalues—forthevisiblecolumns—thisstatementwouldwork.
Now,notethesecondINSERTstatement.Itdiffersfromthefirstonlyinthat
itspecifiesallthreecolumnnames,includingtheinvisiblecolumnname.This
secondINSERTexecutessuccessfully.
So,canyouinsertvaluesintotheinvisiblecolumnofatable?Theansweris
yes,providedthatyouspecifythecolumnnamesintheINSERTstatement.
Takenoteofthispattern;youwillseeitagain,asweturnourattentiontothe
useofinvisiblecolumnsandVIEWobjects.
InvisibleColumnsandViews
Rememberthataviewisanobjectbuiltonaqueryofoneormoretables.When
wespeakofaviewandinvisiblecolumns,wearetalkingaboutthecolumnsof
thetableortablesonwhichtheview’squeryisbased.Arethoseinvisible?Ifso,
thereareseveralissuestoaddress.
Asyoujustlearned,ifatablehasoneormorecolumnsspecifiedas
INVISIBLE,thenwildcardreferences(withtheasterisk)inanINSERT
statementwillnotrecognizetheinvisiblecolumn.Thesameissueappliesto
queriesusedtobuildaVIEW.
Forexample,let’sstartwiththesameSHIP_ADMINtablewejustcreated.
Remember,ithasthreecolumns,oneofwhichisinvisible.Italsonowhasone
rowofdata.SeeFigure10-3.
FIGURE10-3 TheSHIP_ADMINtablewithonerowofdata
Let’screatetwodifferentviewsontheSHIP_ADMINtable.
Forthefirstview,ourquerywilluseawildcardreferencetoallthetable’s
columns.Remember,thequeriedSHIP_ADMINtablehastwovisiblecolumns
andoneinvisiblecolumn.
SeeFigure10-4.Wecreatetheviewandthendescribeit.Theinvisible
columnisomittedfromtheview’sstructure.
FIGURE10-4 VW_SHIP_ADMINviewwithSELECT*andSELECT
columnlist
Next,weexecuteaSELECTstatementthatalsousesawildcardreferenceto
“all”thecolumns,andonlyvisiblecolumnswillbereturnedasoutput.The
invisiblecolumnisnotaccessible.EvenifweexecuteaSELECTthatnamesall
thecolumns,includinganyinvisiblecolumns,theinvisiblecolumnisnot
accessible;Figure10-4showstheresultsoftheattempt.
Eventhoughtheviewisbuiltonaquerythatusesawildcardreferenceto
“all”columnsintheSHIP_ADMINtable,onlythevisiblecolumnsareincluded.
Foroursecondview,let’sspecifytheinvisiblecolumnintheCREATEVIEW
statement’squery;seeFigure10-5.
FIGURE10-5 VW2_SHIP_ADMINviewwithSELECT*andSELECT
columnlist
Nowweseeadifferentoutcome.Thissecondviewisbasedonaquerythat
specifiesaninvisiblecolumnbyname.Inthissituation,theviewwillseethe
invisiblecolumns.Theinvisiblecolumnisreflectedintheview’sstructure.
ASELECTstatementonthissecondviewwillreturnvaluesfromthe
invisiblecolumn,whethertheSELECTstatement(ontheview)usesawildcard
referenceornot,asshowninFigure10-5.Thedifferenceistheuseofthe
wildcardinthequerythatisusedtocreatetheviewinthefirstplace.
Ifthequeryusedtocreatetheviewemploysawildcardcolumnreference,
invisiblecolumnsareignored,buttheviewwillbecreatedsuccessfully;itwill
simplyomittheinvisiblecolumn.
However,whenworkingwithatableconsistingofoneormoreinvisible
columns,itispossibletocreateaviewthatincludesthoseinvisiblecolumnsby
specifyingtheinvisiblecolumnnamesinthequeryusedtocreatetheview.The
resultisaviewwherethecolumnthatisinvisibleinthetableisvisibletothe
view,bothintheview’sstructureandininteractionswiththeviewitself,suchas
SELECTstatements.
CERTIFICATIONOBJECTIVE10.03
Create,Maintain,andUseSequences
Asequenceisanobjectthatispredominantlyusedforonepurpose:togenerate
dataforprimarykeycolumnsintables.Whilethatistheprimarypurposeofa
sequence,there’snothinginherentinthestructureofasequencetolimityouto
suchapurpose.Youcanuseasequencetogeneratenumericsequencesforany
reason.Butallasequencedoesisissuesequentiallyincreasing(ordecreasing)
numbers,accordingtotherulesyoudefinewhenyoucreatethesequence.
CreatingandDroppingSequences
Here’sasampleoftheSQLstatementtocreateasequence:
Thisexampleisacompletestatementandrepresentsthesimplestformofa
CREATESEQUENCEstatement.Thesyntaxisasfollows:
TherequiredCREATESEQUENCEkeywords
Therequirednameofthesequencethatyouspecify,accordingtothe
rulesofnamingdatabaseobjects
Notethatnothinginthecodetiesittoaparticulartableorotherdatabase
object—nothing,thatis,otherthanperhapsthechoiceofthename,whichisa
namingconventionweusebutisnotrequired.
Here’sthecompletesyntaxforasequence:
Thereareseveralsequence_optionsthatcaneachbespecified,separatedby
spaces,asdesired.Sequencescanbesettostartatanynumberandincrement—
ordecrement—byanynumber.Theycansequentiallygeneratewithoutceasing
orbegivenarangewithinwhichtheycontinuouslygeneratenewnumbers.They
canbegivenafixedseriesofnumberstogenerate,afterwhichtheycease
generatingnumbers.
Thesequenceoptionsincludethefollowing:
INCREMENTBYintegerEachnewsequencenumberrequested
willincrementbythisnumber.Anegativenumberindicatesthesequence
willdescend.Ifomitted,theincrementdefaultsto1.
STARTWITHintegerThisspecifiesthefirstnumberthatwillstart
thesequence.Ifomitted,STARTWITHdefaultstoMINVALUE(which
wewilldiscussinabit)forascendingsequencesorMAXVALUEfor
descendingsequences,unlessNOMINVALUEorNOMAXVALUEare
specifiedeitherexplicitlyorimplicitly(bydefault),inwhichcase
STARTWITHdefaultsto1.
MAXVALUEintegerThisspecifiesthemaximumnumberforthe
sequence.Ifomitted,thenNOMAXVALUEisassumed.
NOMAXVALUEThisspecifiesthatthereisnoMAXVALUE
specified.
MINVALUEintegerThisspecifiestheminimumnumberforthe
sequence.Ifomitted,NOMINVALUEisassumed,unlessaMINVALUE
isrequiredbythepresenceofCYCLE,inwhichcasethedefaultis1.
NOMINVALUEThisspecifiesthatthereisnoMINVALUE
specified.
CYCLEWhenthesequencegeneratorreachesoneendofitsrange,
restartattheotherend.Inotherwords,inanascendingsequence,once
thegeneratedvaluereachestheMAXVALUE,thenextnumbergenerated
willbetheMINVALUE.Inadescendingsequence,oncethegenerated
valuereachestheMINVALUE,thenumbergeneratedwillbethe
MAXVALUE.
NOCYCLEWhenthesequencegeneratorreachestheendofits
range,stopgeneratingnumbers.NOCYCLEisthedefault.Ifnorangeis
specified,NOCYCLEhasnoeffect.
Youcanalsodropasequence,likeso:
Here’sanotherexampleoftheCREATESEQUENCEstatement:
ThisSQLstatementperformsthesametaskastheearlierCREATE
SEQUENCEstatementyousaw.Thisexampleexplicitlyspecifiesthedefault
features.Youcanadjustthosedefaultsifyouwant,likethis:
Thisstatementwillstartwiththenumber10andincrementeachsuccessive
numberby5.
UsingSequences
Nowthatwe’vecreatedasequence,whatdowedowithit?Here’sanexample:
InthissampleINSERTstatement,weinsertarowintotheORDERStable
thatconsistsofthreevalues.ThefirstvalueisSEQ_ORDER_ID.NEXTVAL.
ThisreferenceistothesequencegeneratorSEQ_ORDER_IDalongwithits
pseudocolumnNEXTVAL,whichperformsthefollowingtwotasks:
Advancesthesequencegeneratortothenextavailablenumber
Returnsthatvalue
IfthesequencegeneratorSEQ_ORDER_IDhadjustbeencreatedandifit
wascreatedwiththedefaultvaluesfortheSTARTWITHandINCREMENT
BY,thentheinitialcalltoNEXTVALwillobtainthestartingvalueof1.Ifthe
nextcalltothesequencegeneratorisalsoacalltotheNEXTVAL
pseudocolumn,thenitwillbeadvancedagainby1toavalueof2.
Allsequencegeneratorshavetwopseudocolumns:
NEXTVALThisincrementsthesequencetothenextnumber,
accordingtothesequence’soriginalCREATESEQUENCEdirectives.It
alsoreturnsthenewlyincrementednumber.
CURRVALThisdisplaysthecurrentnumberthatthesequenceis
holding.However,thiscallisvalidonlyfromwithinasessioninwhich
theNEXTVALpseudocolumnhasalreadybeeninvoked.Youcannotuse
CURRVALinyourinitialcalltoanygivensequencegeneratorwithina
session.
TheadvantagetoCURRVALbecomesapparentwhenworkingwithasetof
tablesthatinvolvePRIMARYKEYandFOREIGNKEYrelationships.Consider
theentity-relationshipdiagram(ERD)inFigure10-6.
FIGURE10-6 ERDdiagramfortheCRUISE_CUSTOMERSand
CRUISE_ORDERStables
Let’screateacoupleofsequencegeneratorsforusewiththesetables.
Nowlet’sinsertsomenewrowsintothesetables.
Therearethreecallstooursequencegeneratorsintheprecedingcode:
Inline4wecallSEQ_CRUISE_CUSTOMER_ID.NEXTVALto
generateanewprimarykey.
Attheendofline9wecallthesamesequencegeneratorwiththe
CURRVALpseudocolumn.Thisdirectsthesequencegeneratortousethe
samevaluethatwasjustassignedintheline4call,ensuringthatour
PRIMARYKEY–FOREIGNKEYrelationshipbetweenthesetwotables
willberespected.
Atthebeginningofline9,wecall
SEQ_CRUISE_ORDER_ID.NEXTVALtogenerateanewprimarykey
fortheCRUISE_ORDERStable.
TheresultoftheseINSERTstatementsandusesofthesequencegenerators
helpstoensurethatwecanjoinourtablestoproducevalidandmeaningful
output.
Hereareafewimportantpointstokeepinmindaboutsequences:
YoucannotinvokeCURRVALinyourfirstreferencetoasequence
withinagivensession.NEXTVALmustbethefirstreference.
IfyouattempttoexecuteastatementsuchasanINSERTthat
includesthesequencereferenceNEXTVAL,thesequencegeneratorwill
advancetothenextnumbereveniftheINSERTstatementfails.
YoucannotinvokeCURRVALorNEXTVALintheDEFAULT
clauseofaCREATETABLEorALTERTABLEstatement.
YoucannotinvokeCURRVALorNEXTVALinthesubqueryofa
CREATEVIEWstatementorofaSELECT,UPDATE,orDELETE
statement.
InaSELECTstatement,youcannotcombineCURRVALor
NEXTVALwithaDISTINCToperator.
YoucannotinvokeCURRVALorNEXTVALintheWHEREclause
ofaSELECTstatement.
YoucannotuseCURRVALorNEXTVALinaCHECKconstraint.
YoucannotcombineCURRVALorNEXTVALwiththeset
operatorsUNION,INTERSECT,andMINUS.
YoucancallasequencepseudocolumnfromanywherewithinaSQL
statementthatyoucanuseanyexpression.
Thatlastpointisimportant—areferencetoasequencemustincludeits
pseudocolumn,andsuchareferenceisconsideredtobeavalidexpressionora
componentofanexpression.So,assumingyou’reworkingwithatable
PROJECTSthathasacolumnPROJECT_COST,youcouldinvokeasequence
SEQ_PROJ_COSTlikethis:
Thatisvalidsyntax.Whetherit’susefulornotisuptoyourbusinessrules.
ButSQLrecognizesthissyntaxandwillexecuteitsuccessfully.
Thebottomlineisthatreferencestothepseudocolumnsofsequencesare
validexpressions.
AnySQLstatementthatisexecutedwithacalltoasequence’s
NEXTVALpseudocolumnwilladvancethesequence,eveniftheSQL
statementfails.NotevenaROLLBACKwillresetasequence;eventhe
valueofthesequencegenerator’sCURRVALwillremainthesameafter
aROLLBACK.
CERTIFICATIONOBJECTIVE10.04
CreateandMaintainIndexesIncludingInvisible
IndexesandMultipleIndexesontheSameColumns
Anindexisanobjectyoucancreateinthedatabase,anditstoresasubsetofdata
fromatable.Theindexisitsownobject;itisstoredseparatelyfromthetable,
anditisconstantlymaintainedwitheveryINSERT,UPDATE,andDELETE
performedonthetable.Forexample,ifyouhaveatableSHIPSwithacolumn
CAPACITYandyoucreateanindexforthetableusingthecolumnCAPACITY,
thenboththetableandtheindexwillstore—separately—CAPACITYvalues.
EverytimeaDMLstatementchangesavalueintheSHIPStable’sCAPACITY
column,theindexwillalsobechangedaccordingly.
Thepurposeofanindexistosupportfasterqueriesonatable.Theindex
valuesarepresorted.Anindexcanbecreatedforonlyonetable,anditcanbe
createdononeormoreofthattable’scolumns—columnsthatyoudesignate.For
eachcolumn,theindexalsostorestheaddressofthatdatafromthesourcetable.
SQLcanthenusetheindexobjecttospeedupthequeryingofWHEREand
ORDERBYclauses.Forexample,ifaWHEREclausereferencesanyindexed
column(orcolumns),thenSQL(ormorespecifically,theoptimizer)will
automaticallyconsidertheindexasitdeterminestheoptimalquerystrategyfor
theSQLstatement.Theresultisthatqueriesmaybesignificantlyfaster,
dependingontheamountofdatainvolvedanddependingonthenumberof
indexesthatmaybeappliedtoatable.
Notethatwesaythatanindex“may”speedupperformance.SQLstatement
processingisperformedbyabuilt-infunctionknownastheOracleDatabase
optimizer.TheoptimizerdetermineshowbesttoprocessanygivenSQL
statementaccordingtoacombinationoffactors.Twoofthosefactorsarethe
presenceandstructureofindexes,butthereareotherfactorsaswell.AsSQL
developers,itisourjobtocreatethenecessaryresourcesfortheoptimizertodo
itsjob,recognizingthattheoptimizer,intheend,makesthefinaldecisionand
processesthestatement.Iwillspeakmoreabouttheoptimizerinafew
paragraphs.
YoucannotcreateanindexoncolumnsofLOBorRAWdatatypes.Given
that,youcancreateanynumberofindexes,limitedonlywithregardtocertain
aspectsofcolumnsets,whichIwillexplaininabit.However,asyouadd
indexes,you’lleventuallyreachapointofdiminishingreturns—eachindex
addedtoatablecanpotentiallyincreasetheworkloadonfutureINSERT,
UPDATE,andDELETEstatements.Butlet’sholdoffonthatdiscussionfora
bit.Fornow,wejustwanttosaythattheSQLsystemwillletyoucreateasmany
indexesasyouwant.
RememberthattheWHEREclausecanappearinaSELECT,UPDATE,or
DELETEstatement.AlsonotethataSELECTstatement,intheformofa
subquery,mayappearwithinanINSERTstatementoravarietyofstatements
thatcreateobjectssuchasTABLEorVIEW.AnINDEXobjectcanpotentially
benefitanyofthesesituations.
TheOracleDatabaseOptimizer
ThereasonwecreateanindexonatableissothattheOracleDatabaseoptimizer
canprocessSQLstatementsonthattablefaster.Butwhatistheoptimizer?
Theoptimizerisabuilt-infeatureintheOracledatabasethatiscrucialtothe
processingofeverySQLstatementissuedagainstthedatabase.EachSQL
statementhasthepotentialforgreatcomplexity,withinstructionstojoin,filter,
andtransformdatainavarietyofways.Theamountofdatatobeprocessedcan
varydramatically.Thiscomplexityintroducesahostofchallenges.The
optimizerisdesignedtohandlethosechallenges,anditperformsverywell.
Forexample,ifyouwanttoexecuteatypicalSQLstatementwithsomescalar
functionsandaWHEREclause,thenitwouldbeidealiftheWHEREclause
executesfirstandthescalarfunctionsareappliedlater.Why?Becausethe
WHEREclausecanfirstreducethenumberofrowstobeprocessed,leaving
onlythoserowsnecessaryonwhichtoapplythescalarfunctions.The
responsibilityofensuringthatsuchprioritizationisfactoredintoSQLprocessing
isbuiltintothedatabase,andtheoptimizeristhefeaturethatanalyzestheSQL
statementandestablishesanexecutionplanforprocessingthestatement.This
executionplanisaccessibletoyouviatheEXPLAINPLANstatement,whichis
notasubjectoftheexam,butisworthinvestigating.Ifyouareinterestedin
learningmoreabouttheoptimizer,lookfordocumentationontheEXPLAIN
PLANstatement,andyouwilllearnagreatdeal.
Forourpurposes,itisenoughtoknowthattheoptimizeristhebuilt-in
functionthatprocessesSQLstatementsandthat,aspartofthatprocessing,the
optimizeristhedecisionpointwherethedatabasedetermineswhetherandhow
toincorporateanindexintheexecutionofanygivenSQLstatement.
Yourgoalwhenbuildingindexesistocreateasetofusefulobjectstobe
leveragedbytheoptimizerinitsconstructionoftheoptimalexecutionplanfor
processingfutureSQLstatements,whateverthosemaybe.Yourjobasa
developeristoanticipatethetypicalusageofyourtableandthencreateindexes
thattheoptimizerwillbeabletousetobestexecutethosefutureSQL
statements.Understandinghowindexesworkisanecessarysteptoward
achievingthatgoal—andperformingwellontheexam.
Wewillrevisitthetopicoftheoptimizeraswediscussindexesinthe
followingsections.
ImplicitIndexCreation
IfyoucreateaconstraintonatablethatisoftypePRIMARYKEYorUNIQUE,
thenaspartofthecreationoftheconstraint,SQLwillautomaticallycreatean
indextosupportthatconstraintonthecolumnorcolumns,ifsuchanindexdoes
notalreadyexist.
Forexample,considerthefollowingSQLstatement:
ThisstatementwillcreatethetableSEMINARS,twoCONSTRAINTs,and
twoINDEXobjects.TheseINDEXobjectswillbenamedautomaticallybythe
SQLsystem.Laterinthissection,you’llseehowtomanuallycreatethese
indexes.Alsolater,inChapter12,you’llseehowyoucanquerythedata
dictionarytoseetheseimplicitlycreatedindexes,butfornow,here’saqueryon
thedatadictionarythatcanconfirmtheircreation:
Intheoutputofthisexample,youcanseethatthesystemassignedthenames
SYS_C009932andSYS_C009931totheindexes.
Asanalternative,youcanqueryelsewhereinthedatadictionarytoseethe
columnsthatareinvolvedintheindexes:
Intheseexamples,weappliedeachconstrainttoasinglecolumn.Hadwe
createdacompositePRIMARYKEYoracompositeUNIQUEconstraint,then
allcolumnsinvolvedintheconstraintwouldhavealsobeenindexedasa
compositeindex,andthecolumnnameswouldbelistedintheoutputofthis
secondqueryonthedatadictionary.
SingleColumn
HereisanexampleofaSQLstatementthatcreatesasingle-columnindex:
ThisSQLstatementcreatesasimpleindexcalledIX_INV_INVOICE_DATE,
onthecolumnINVOICE_DATE,inthetableINVOICES.Thiswillresultinan
objectthatsortsvaluesintheINVOICE_DATEcolumnoftheINVOICEStable.
Oncesuchanindexexists,queriesthatreferenceINVOICE_DATEinthe
WHEREclausemayreturnresultssignificantlyfaster.
Aswestatedearlier,theindexmayormaynotspeedtheperformanceofthis
SELECTstatement.Thedecisionisuptotheoptimizer.Rememberthatthe
optimizeranalyzesallSQLstatementstodeterminethebestcourseofactionfor
executingthestatement.ForanySQLstatementthatinvolvesatablesearchor
scan,suchasaquerywithaWHEREclauseoranORDERBYclause,the
optimizerwillconsiderwhetheranavailableindexonthequeriedtablewill
contributetowardfasterperformance.Theexistenceofanindexdoesnot
guaranteeitsuse;theoptimizerdetermineswhetherornottheindexwillbe
helpful.
Inourexample,theSELECTstatementagainsttheINVOICEStableincludes
aWHEREclausethatspecifiestheINVOICE_DATEcolumn.If,forexample,
alltherowsintheINVOICEStablehavethesamevalueforINVOICE_DATE,
theindexmaynotbeused—itmightcontributenoperformancebenefit.Onthe
otherhand,ifeachrowhappenstohaveauniqueINVOICE_DATEvalue,then
theindexmostlikelywillbeused.Thedifferencehastodowiththeconceptof
selectivity.Ifacolumntendstoincludedatathatislessrepetitiveandmore
unique,itissaidtohaveahigherdegreeofselectivity,andanindexonsucha
columnwouldbeattractivetotheoptimizer.Butifdatainagivencolumnis
relativelyrepetitive,itissaidtohavealowerdegreeofselectivityandwillbe
lesslikelytobeincludedinanindex.Theoptimizeristhefeaturethatmakesthat
decision.
InthesampleSELECTyoujustlookedat,iftheSQLoptimizerdeterminesto
usetheindex,andifevidenceisfoundintheindexthatarowofINVOICES
matchesthedesiredresult,theindexpointsthequerydirectlytotherow(or
rows)inthetable,andtheresultisachieved—muchfasterthanotherwisewould
haveoccurred.
Usage
TheOracleDatabaseoptimizerwillconsidertheuseofanindexinanyquery
thatspecifiesanindexedcolumnintheWHEREclauseorORDERBYclause,
dependingonhowtheindexcolumnisspecified.Therearesomeguidelinesfor
structuringyourSQLstatementsothattheoptimizerwillbemorelikelytoapply
theindex.Theguidelinesareasfollows:
Forbestresults,theindexedcolumnshouldbespecifiedina
comparisonofequality.
A“greaterthan”orsomeothercomparisonmaywork.
A“notequals”willnotinvokeanindex.
TheLIKEcomparisonmayinvokeanindexaslongasthewildcard
characterisnotintheleadingposition.Inotherwords,theexpression
LIKE'%SMITH'willnotinvokeanindex,butLIKE'SMITH%'may
invokeanindex.
Afunctiononacolumnwillpreventtheuseofanindex—unlessthe
indexisafunction-basedindex.(Function-basedindexesarenot
addressedintheexam.)
Furthermore,certaintypesofautomateddatatypeconversionswilleliminate
theindex.
Thedesignofanindexingschemeissomethingthatfallsunderthegeneral
categoryofperformancetuning.Thereisnosinglecomprehensiveapproachto
indexdesignforagivendatabaseapplication.Theapproachyoutakedepends
uponavarietyoffactors,includingtheamountandselectivityofdataexpected
andtheanticipatedfrequencyandtypesofthevariousSQLstatementsyou
intendtoexecuteonthedatabase.Asystemthatwillnotbequeriedmuchbut
willbeheavilypopulatedwithINSERTstatementswouldnotbenefitmuchfrom
indexes,sinceeachINSERTwillbeslowerasitupdatesthetablesandall
associatedINDEXobjects.Onsuchadatabase,infrequentqueriesmaynot
justifythecreationofINDEXobjectsandtheirassociatedperformancetrade-off.
Ontheotherhand,anapplicationthatinvolvesagreatdealofupdatingvia
UPDATEstatementswillbenefitfromthoughtfullydesignedindexes.
Theoptimizerwilldoeverythingitcantouseanindexforagiven
querywheretheindexisbeneficial.Itavoidsanindexonlyifitsuse
wouldbedetrimental,suchasinthecaseofatablewithalargenumber
ofrowswithanindexedcolumnthathaslowselectivity.
Anoldruleofthumbrecommendsnomorethanfiveindexesontheaverage
tableinatransaction-basedapplication.Butthat—again—dependsonintended
usage.I’veimplementedapplicationsinwhichsometableshavehadasmanyas
nineindexesapplied,andtheresultwaspositive.Thatwasasituationwhere
queriesonthetablewerenotusingprimarykeyssomuchbutinsteaduseda
varietyoftext-basedlookups.
Performancetuningisthesubjectofmanybooksandisatopicthatgoesfar
beyondthisbook—andtheexam.Fortheexam,it’simportanttounderstandthe
syntaxofcreatingindexes—bothsinglecolumnandcomposites—andtheir
intendedpurposeandgeneralusage.
Maintenance
IndexesareautomaticallymaintainedinthebackgroundbySQLandrequireno
effortonyourparttokeeptheirlookupdataconsistentwiththetable.However,
notethatthismeanseachfutureINSERT,UPDATE,andDELETEstatementyou
issuewillhavetoworkharder.EachDMLstatementthatmodifiesdatainatable
thatisindexedwillalsoperformindexmaintenanceasrequired.Inotherwords,
eachindexyouaddtoatableputsmoreworkloadoneachDMLstatementthat
affectsindexeddata.
Buildtheindexesthatcontributetoyourapplication’soverallperformance,
anddon’tbuildindexesunlesstheyarenecessary.Plantoperiodicallyreview
yourindexstructurestodeterminewhichindexesmaynolongerbeneeded;an
existingapplicationburdenedwithunnecessaryindexesmaybenefitfrom
removingindexes.
Thereareanumberofperformancetuningtoolsandtechniquestoassistwith
designingindexes.OneistheEXPLAINPLANfeature,whichrevealsthe
internalworkingsofagivenSQLstatementandprovidesinsightintohowSQL
willoptimizeagivenstatement,includingtheuseofindexesforagivenSQL
statement.It’sagreattool.Itisnotacertificationobjectivebutyoumaywant
toexploreoneoftheotherbooksintheOraclePresslinetolearnmoreabout
it.
Composite
Acompositeindexisanindexthatisbuiltontwoormorecolumnsofatable,
likethis:
Inthisexample,theresultisonesingleindexthatcombinesbothcolumns.A
compositeindexcopiesandsortsdatafrombothcolumnsintothecomposite
indexobject.Itsdataissortedfirstbythefirstpositioncolumn,secondbythe
secondpositioncolumn,andsoon,forallthecolumnsthatformthecomposite
index.
AWHEREclausethatreferencesallthecolumnsinacompositeindexwill
causetheSQLoptimizertoconsiderusingthecompositeindexforprocessing
thequery.Forexample,thisquerywouldencouragetheoptimizertousethe
compositeindexIX_INV_INVOICE_VENDOR_IDtomaximumadvantage:
NotethattheWHEREclauseintheprecedingexamplereferencesthe
columnsthatmakeupthecompositeINDEXobject.Next,considerthisquery:
Thisquery,whichreferencesthefirstcolumninthecompositeindex,also
invokestheindex.Thereasonisthatthecompositeindexinternallysortsdataby
thefirstcolumnfirstandthesecondcolumnsecond.Giventhat,theinternally
copiedandsorteddatafromtheindexedtableiscomparableinstructuretoa
single-columnindexbasedon—inthisexample—theVENDOR_IDcolumn.
Butnowconsiderthisquery:
Thisqueryreferencesthesecondcolumninthecompositequery,butnotthe
first.Thecompositeindexstructureisprimarilysortedonthefirstcolumnofits
structure,whichintheexamplewe’reworkingwith,isVENDOR_ID.Butthis
querydoesnotreferenceVENDOR_ID.Nevertheless,theSQLoptimizermay
stillconsiderapplyingtheindexbecauseofafeatureknownasskipscanning.
SkipScanning
Thankstoskipscanning,aWHEREclausethatreferencesanycolumnswithina
compositeindexmayinvoketheindexinitsprocessing.However,the
performancebenefitisnotidentical.Inskipscanning,SQLtreatsacomposite
indexasacombinationofseveralindexes.Howmany?Thatdependsonthe
levelofselectivitythatresultsintheindexingofthefirst(orleading)columnof
theINDEXobject.Ifthefirstcolumncontainsonlyafewuniquevalueswithina
largenumberofrows,theindex—ifused—maybeappliedarelativelyfew
numberoftimes.Ontheotherhand,iftheleadingcolumncontainsdatathatis
relativelyuniqueacrosstherows,thentheindex—ifused—maybereviewed
frequently.Inotherwords,askipscanwilldoanindexscanonceforeach
uniquevalueinthefirstcolumn.Thisisn’tquiteasbeneficialasasimpleone-
columnindex,anditsbenefitvaries,dependingontheuniquenessofvaluesin
thefirstcolumn.ButtheWHEREclausegainssomebenefitanyway.
ThepointisthataWHEREclausethatreferencessome,butnotall,ofthe
columnsinacompositeindexmayinvoketheindex,eveniftheleadingcolumn
isnotreferencedintheWHEREclause.However,includingtheleadingcolumn
mayresultinamoreefficientapplicationofthecompositeindexand,therefore,
afasterresulttothequery.
Unique
Auniqueindexisonethathelpsensurethatacolumninatablewillcontain
uniqueinformation.Thesyntaxforcreatingauniqueindexisasfollows:
ThisSQLstatementcreatesanindexthatwillensurethatdataenteredintothe
SSNcolumnofthetableEMPLOYEESisunique.
ThisisdifferentfromtheUNIQUEconstraintthatyoucanapplytoacolumn
onatable.However,notethatifyoucreateaPRIMARYKEYorUNIQUE
constraintonatable,auniqueindexwillautomaticallybecreatedalongwiththe
constraint.NotethattheUNIQUEconstraintismoreself-documentingwithin
thedatabase.Thatbeingsaid,OracleCorporationformallyrecommendscreating
uniqueindexestoenforceuniquenessinacolumnforbetterresultsinquery
performance.
Inpractice,avoidusingSocialSecuritynumbersas“unique”values.Inthe
realworld,itturnsoutthatSocialSecuritynumbers,whichareissuedbythe
U.S.federalgovernment,arenotnecessarilyunique,somethingthatmany
databaseprofessionalshavelearnedthehardway.OnceanSSNisnolonger
needed,itcanbere-issuedtosomeoneelse.Therearemanyclassicstoriesout
thereaboutapplicationdatabasesthatstartedemployingSSNsasprimarykeys
onatable,onlyfortheproductiondatabasetoencounterreal-worlddata
wheremorethanonepersonhasbeenissuedthesameSSN.It’sagreat
argumentforalwaysmakingyourprimarykeyswithsequence-generated
valuesandalessonthatSSNsintherealworldaren’treallyunique.Anddon’t
evengetmestartedontheissueofsecurityandthelawwithregardtoSSNs.
Unlessyouaresupportingagovernment-mandatedrequirementtoreport
somethingwithregardtoSocialSecuritytaxesorpayments,itisbesttoavoid
storingtheSSNatall—forsecurity,iffornootherreason.
Dropping
YoucandropanindexwiththeDROPINDEXstatement.Forexample,let’s
dropourIX_INV_INVOICE_DATEindex:
Notethatifyoudropatableuponwhichanindexisbased,theindexis
automaticallydropped.Ifyoure-createthetable,youneedtore-createtheindex.
VisibleandInvisibleIndexes
Anindexcanbemadetobevisibleorinvisibletotheoptimizer.Onlyavisible
indexisrecognizedbytheoptimizerand,therefore,availablefortheoptimizerto
useintheprocessingofaSQLstatement.Ifanindexismadetobeinvisible,the
optimizerwillomittheindexfromconsiderationwhencreatingtheexecution
planfortheSQLstatement.
Anindexisvisiblebydefault.Youmustexplicitlydeclareanindextobe
invisibleifthatisyourpreference.
(Thereisaninitializationparameter
OPTIMIZER_USER_INVISIBLE_INDEXESthatcanbesettoTRUEatthe
systemorsessionlevel.IfTRUE,theoptimizerwillconsiderallindexes
regardlessofvisibility—orinvisibility.However,initializationparametersand
thestatementstosetthemarenotthesubjectoftheexam.)
WhyUseanInvisibleIndex?
So,whywouldyoucreateanindexwhosesolepurposeinlifeistospeed
processingandthenrenderitinvisiblesothatitcannotbeusedtospeed
processing?Thefollowingsectionshighlightsomereasons.
IndexTestingonaTableMatureapplicationsmayhaveanumberofdifferent
indexesappliedtoatable.Alloftheseindexesrequiremaintenance.Thatmeans
everytimeanINSERT,UDPATE,orDELETEisexecutedonthetable,allof
thattable’sindexesalsohavetobeupdated(or“maintained”)withthesamedata
forconsistency.Therequisiteprocessingcarriesapriceintermsofperformance
andstorage.Isthisoverheadworthit?Theanswerdependsonwhetherthe
indexesareuseful.Onewaytofindoutistotemporarilychangeanindexto
invisibleandthenrunyourqueries.Theinvisibleindexwillbeignoredbythe
optimizer,givingyouanopportunitytoassesstheimpactofomittingtheindex
withouthavingtodropitaltogether.Isperformancedegradedwhenthe
optimizerignorestheindex?Ifnot,thentheindexisnotimprovingperformance.
YoucouldthenchoosetodroptheindexaltogetherandreducetheDML
maintenancecostassociatedwithitsexistence.
VISIBLE,INVISIBLE
ThekeywordsVISIBLEandINVISIBLEspecifythevisibilityofanindexinthe
CREATEINDEXstatement.Forexample,consideratablePORTS.
ThefollowingcodecreatesaninvisibleindexonthePORT_NAMEcolumn:
Alternatively,hereishowtocreatethesameindexasavisibleindex:
VISIBLEisthedefault.IfaCREATEINDEXstatementomitsthekeywords
VISIBLEorINVISIBLE,avisibleindexiswhatwillbecreated.
Iftheindexexistsandyouwanttoalterittobeinvisible,usethis:
Toalteranindexandmakeitvisible,usethis:
Iftheindexisalreadyinvisible(orvisible)andyousubmitanALTER
statementto“change”ittowhatitalreadyhappenstobe,thestatementwill
executesuccessfullywithnocomplaintorindicationthattheALTERwas
unnecessary.
ThedatadictionaryviewUSER_INDEXESincludesacolumnVISIBILITY
thatreportswhetheranindexiscurrentlyvisibleorinvisible.
Rememberthatanindexisvisiblebydefault,soifyouomittheVISIBLE
keyword,youstillgetavisibleindexatthetimeofindexcreation.
Thedatadictionaryusesuppercaseletterstorecordthenamesofall
tables,columns,views,andotherobjectsyoucreate,unlessyouenclose
thenameindoublequoteswhenyoucreatetheobject.
IndexAlternativesontheSameColumnSet
Youcancreateanumberofindexesonagiventable.However,foragiventable,
youarelimitedwhencreatingmultipleindexesonaparticularcolumnset,and
by“set,”wemeanasetofoneormorecolumnsforthattable.Tocreatemultiple
indexesonagivencolumnset,theindexesmustdifferwithregardtooneofaset
ofspecificcharacteristics.However,whileyoucancreatemultipleindexes
withintheselimitations,onlyoneofthoseindexobjectsmaybevisibleatany
pointintime.
Tocreatemultipleindexesonagivencolumnset,theindexesmustvarywith
regardtooneofthefollowingcharacteristics:
DifferentuniquenessYoucancreateauniqueindexonacolumnset
andanonuniqueindexonthesamecolumnset.
DifferenttypesTheindexeswehavelookedatsofarhavebeen,by
default,B-treeindexes.Youmayalsocreateabitmapindexbyadding
thekeywordBITMAPpriortothewordINDEX,asinCREATE
BITMAPINDEX,followedbytheappropriatekeywordstocompletethe
indexspecification.AB-treeindexisidealfortransactionalapplications
wherethefullsetofDMLstatementswillbecommonlyutilized.A
bitmapindexisoftenusedmoreinapplicationswhereheavieruseof
SELECTstatementsisanticipated.Forourimmediatepurposes,thepoint
isthatyoucancreateindexesonthesamecolumnsetiftheydifferwith
regardtoindextype.
DifferentpartitioningMultipleindexesmayexistonthesame
columnsetprovidedthatcertainpartitioningcharacteristicsaredifferent:
Partitioningtype(hashorrange)
Localversusglobalpartitioning
Partitioningversusnonpartitioning
Forthepurposeoftheexam,thepointyouneedtorememberisthatwhile
thesemultipleindexesonagivencolumnsetmayexistsimultaneously,onlyone
maybevisibleinthedatabaseatanygivenpointintime.Forexample,when
creatingmultipleindexesonagivencolumnset,anypreexistingindexesonthe
samecolumnsetmustalreadybeinvisible,orelsetheattempttocreateanew
visibleindexonthesamecolumnsetwillfail.
Onereasonfortogglingindexesbetweenvisibilityandinvisibilityistotune
applicationperformanceforvariouspurposes.Rememberthattheexistenceof
anindexdoesnotguaranteetheoptimizerwillutilizetheindex.However,
makinganindexinvisibleguaranteestheoptimizerwillignoretheindex.
Considerthesectionwejustreviewed.Youmightfinditadvantageoustotry
differenttypesofindexesonagiventypetooptimizeperformancefor
transaction-basedapplicationsversesheavyqueryingandanalytics.Thedetails
ofsuchconsiderationsarebeyondthescopeoftheexam;forourpurposes,itis
helpfultorecognizethatapplicationtuningisausefulreasontoemployinvisible
indexes.
Butrememberthatevenifanindexisinvisible,itisstillmaintained.Any
INSERT,UPDATE,orDELETEstatementprocessedonthetablewillalso
updatetheindex,whichisaperformancecost.Thisistruewhetheranindexis
visibleorinvisible.
CERTIFICATIONOBJECTIVE10.05
PerformFlashbackOperations
Oracle’sFlashbackoperationsconsistofanumberoffeaturesthatenableaccess
todatabasedataandobjectsintheirpaststate,includingobjectsthathave
alreadybeendropped.ThesefeatureshavenameslikeFlashbackTable,
FlashbackDrop,FlashbackQuery,andmore.ThetopicofOracle’sFlashback
operationsisquiteinvolved.Developingacompleteunderstandingofallofwhat
OracleofferswithFlashbackoperationsrequiresmorestudythanwhatis
requiredtopasstheexam.We’llprovidesomebackgroundinformationtogive
youanideaofwhatwearetalkingabout,andthenwe’lldiscussthespecific
Flashbackoperationsthatarerelevanttotheexamandtothischapter’sfocus,
whichistheoverallsubjectofmanagingschemaobjects.Thisbookwillpresent
whatyouneedtoknowfortheexam,butyoumaywanttocontinueyourstudy
ofFlashbackoperationsbeyondwhatisrequiredfortheexam.
Oneimportantnote:OracleFlashbackoperationsincludesomefeaturesthat
areavailableonlyintheOracleEnterpriseEdition.IfyouareusingtheStandard
Edition,youcanperformFlashbackQuery,butsomeofthefeaturesdescribedin
thissectionrequireEnterpriseEdition—suchasFlashbackTable.
Overview
Oracle’sFlashbackoperationsincludeavarietyofstatementsyoucanuseto
recoverobjectsand/orthedatacontainedwiththem,aswellasdependent
objectsanddata.ThesortoftasksyoucanaccomplishwithFlashback
operationsinclude
Recoveringcompletetablesyoumayhaveinadvertentlydropped
Recoveringdatachangeswithinoneormoretablesresultingfroma
seriesofDMLstatements
Performingdataanalysisondatathat’sbeenchangedoverperiodsof
time
Comparingdatathatexistedatonepointintimewithdatathat
existedatanotherpointintime
Performingqueriesasofapriortimeperiod
Flashbackoperationscansupportmultipleusersessionsgainingaccessto
historicaldatadynamically,onanytable(includingthesametables)atthesame
time,witheachusersessionpotentiallyaccessingdifferentpointsinthehistory
ofthetablesimultaneously,allwhilethedatabaseisupandrunninginfull
operationalmode.
SomeFlashbackoperationsrequirevariousconfigurationsteps;someofthose
configurationscanbeinvolvedandmightrequireinterventionbythedatabase
administrator.Theconfigurationstepsinvolvedcanaffectsystemparameters,
tableclauses,andafeatureofthedatabaseknownastheundosegments,which
haveapurposethatgoesbeyondFlashback.
RecoverDroppedTables
Inthischapter,we’refocusingonlyonmanagingschemaobjects.Withintheset
ofavailableFlashbackoperations,thefeaturethataffectsschemaobjectsasa
wholeistheFLASHBACKTABLEstatement.Thisstatementcanrecovera
previouslydroppedtableyouspecifyfromahistoricalpointintimethatyou
specify.However,therearelimitations.Forexample,youcannotflashbacktoa
pointpriortowhenthetable’sstructuremayhavebeenaltered.
Youcanidentifyapointintimeinavarietyofways:
Immediatelypriortowhenatablewasdropped
AspecifictimeidentifiedbyavalueofdatatypeTIMESTAMP
Aspecifictransactionidentifiedbythesystemchangenumber
(SCN)
Apredeterminedeventidentifiedbyadatabaseobjectknownasthe
RESTOREPOINT
Whenusedtorestoreatable,FLASHBACKTABLErestoresthedropped
tablewitheitheritsoriginalnameoranewnameyouprovidewiththestatement.
Italsorecoversanyindexesonthetableotherthanbitmapjoinindexes.All
constraintsarerecovered,exceptforreferentialintegrityconstraintsthat
referenceothertables—inotherwords,foreignkeyconstraints.Granted
privilegesarealsorecovered.
Note:FLASHBACKTABLEisnotcurrentlyavailableinOracleDatabase
StandardEdition.
ThebeginningsyntaxofaFLASHBACKTABLEstatementisasfollows:
TherequiredkeywordsFLASHBACKTABLE
Oneormoretablenames,separatedbycommas
TherequiredkeywordTO
Inotherwords:
Morethanonetablecanbeincludedinthelist.Additionaltablenamesmust
beseparatedbycommas:
That’sthebeginning.Thereareseveralwaystocompletethisstatement.Here
ishowyoucompleteitifyouwanttorecoveradroppedtable:
TherequiredkeywordsBEFOREDROP
TheoptionalkeywordsRENAMETO,followedbyanewnamefor
thetableifyouwanttorecoverthedroppedtableintoanobjectwitha
differentname
HereisanexampleofaSQLsessionwherewecreateatable,dropit,and
thenuseFLASHBACKTABLEtorestorethedroppedtable:
NotetheFLASHBACKTABLEstatementonline5.Itcombinesthe
beginning:
YoucannotrollbackaFLASHBACKTABLEstatement.
withthefollowing:
BEFOREDROP
ItomitstheoptionalRENAMETOandthenewname.
TheRecycleBin
TheFlashbackDropfeaturerecoverscompletetablesthatarestillretainedinthe
“recyclebin,”anditcandosoinspiteofthefactthatsuchachangeresultsfrom
theDROPTABLEstatement,which,bydefinition,isaDDLstatementand
thereforeinvolvesanimpliedCOMMIT.Inspiteofthis,wecanrecoverthetable
ifitisstillintherecyclebin.
TablesareputintotherecyclebinautomaticallybySQLwheneveraDROP
TABLEstatementisissued.Atable’sdependentobjects,suchasindexes,are
alsoplacedintotherecyclebin,alongwiththetable’sconstraints.
Therecyclebinisnotcountedasspacethatisusedbyagivenuseraccount.
Auseraccount’sdroppedobjectsareretainedinaseparaterecyclebinfor
eachuser.Youcaninspectthecontentsofyourownrecyclebinwiththe
followingquery:
Thatqueryisidenticaltothisone:
RECYCLEBINisasynonymforUSER_RECYCLEBIN.Inotherwords,the
precedingtwoqueriesareidentical.
ThereisaDBA_RECYCLEBIN,whichallowsuseraccountswithdatabase
administrator(DBA)privilegestoseealldroppedobjectsinthedatabase.
Ifyouruseraccounthasprivilegesonanobject,thenyouruseraccountwill
beabletoseetheobjectintherecyclebinintheeventitisdropped.
Youdon’tneedtoinspecttherecyclebinbeforeissuingaFlashback
statement.Butyoumightfindithelpful.
Therecyclebinisaffectedbythe“recyclebin”initializationparameterand
canbeturnedonoroffaccordinglywiththefollowingALTERSESSION
statements:
Bothofthesestatementstakeeffectimmediately.Theinitialstateofthe
recyclebinisdependentonthesettingforrecyclebinintheinitialization
parameterfile,whichiscontrolledbytheDBA.
SettingrecyclebintoONorOFFdetermineswhetheritwillbeused.Any
objectsalreadyintherecyclebinwillremaineitherway.Toremovethem,use
PURGE,whichisexplainedlaterinthischapter.
DependentObjects
Whenatableisrecovered,anyassociateddependentobjectsarealsorecovered,
includingthefollowing:
Indexes,exceptforbitmapjoinindexes
Constraints,butwithlimitations—forexample,restoringdropped
tablesdoesnotrecoverreferentialconstraints,meaningFOREIGNKEY
constraints
Otherobjectsthatwedon’tdiscussinthisbookandthatarenota
subjectoftheexam,suchastriggers
Objectsthathavethesamenamethataredroppedcanallberetrievedwith
Flashbackoperations.Forexample,ifatableVENDORSwasdroppedandthen
re-createdanddropped,therewillbetwoVENDORStablesintherecyclebin.
Thelastonedroppedwillbethefirstoneretrieved.
Objectssuchasindexeswillberecoveredwithsystem-assignednames,not
thenamestheywereoriginallygiven.Youcanrenameeachretrievedobjectwith
theRENAMETOclauseoftheFLASHBACKTABLEstatementasitis
retrieved.Asofthiswriting,renamingretrievedobjectsisbeyondthescopeof
theexam.
StatementExecution
TheFLASHBACKTABLEstatementoperatesasasinglestatement.Ifitfails,
nothinginthestatementsucceeds.Inotherwords,ifthere’sanattempttorestore
threetablesinasinglestatementandthethirdattemptiserroneousforwhatever
reason,noneofthetableswillberestored.
PURGE
ThePURGEstatementpermanentlyremovesthelatestversionofagivenitem
fromtherecyclebin—forexample,topermanentlyremovetheHOUDINItable
fromtherecyclebinsothatitcannotberecovered:
PURGETABLEHOUDINI;
Afterexecutingthisstatement,thetableHOUDINIcannotberecoveredwith
theFLASHBACKTABLEstatementweusedearlier.
NotethatthetablemusthavefirstbeendroppedforPURGEtoexecute
successfully.PurgingmaybeperformedautomaticallybytheOracledatabase’s
ownautomaticspacereclamationoperations.Ifthathappens,thetableisnotin
therecyclebinandcannotberecoveredwithFlashbackoperations.
Youcanalsochoosetopurgetheschema’srecyclebinentirely.
PURGERECYCLEBIN;
Theresultwillbetopurgeallobjectsintheschema’srecyclebin.Topurge
therecyclebinoftheentiredatabase,usethis:
Thepreviousstatementwillexecutesuccessfullyiftheschemahasthe
SYSDBAprivilege.YoucanalsoDESCRIBEandSELECT*fromthe
RECYCLEBINandDBA_RECYCLEBIN.
RecoveringDataWithinExistingTablesoverTime
ThissectiondiscussessomeadditionalwaystocompletetheFLASHBACK
TABLEstatement.
InadditiontoperformingaFlashbackoperationtorestoreadroppedtable,
youcanflashbackanexistingtabletoaspecificpointintime,showingitsstate
priortoanycommittedchangesthatmayhavebeentransactedsincethepointin
timeofinterest.
Thesyntaxcantakeanyofthesethreeforms:
Therecommendedformofthesethreeisthefirst:thesystemchangenumber.
ThisisthemechanismthatisrecommendedbyOracleforidentifyingpointsin
timeinthedatabase.Forexample,let’srevisitourtableHOUDINI.
Let’sreviewtheprecedingcode:
Line3:Wecommitourchangetothetable.
Line4:Thisisastatementthatsuspendsprocessingfor15seconds.
Thechoiceof15secondsisarbitrary;theintenthereistoallowsome
timetopass.
Line6:Wecommitthedeletionofonerow.
Line8andline9:Thisstatementattemptstorestorethetableto
whereitwas20secondsearlier.Notethatline9containsnothingmore
thananexpressionoftheTIMESTAMPdatatype,whichinthiscaseisa
calltotheSYSTIMESTAMPfunctionminusa20-secondinterval.The
SYSTIMESTAMPfunctionreturnsthecurrenttimeasdefinedbythe
Oracledatabaseserver’soperatingsystem.Followingthecalltothe
functionisthesymbolforsubtraction—theminussign—followedbya
literalvaluerepresentingatimeintervalof20seconds.
So,whatistheresultofthisseriesofstatements?Hereitis:
Whatiswronghere?TheproblemisthatthecapabilitytoperformFlashback
operationstorestoreanexistingtabletoanolderstateisnotacapabilitythat
existsbydefault.ItonlyworksontableswheretheROWMOVEMENTfeature
hasbeenenabled.Here’showtocreatethetablewithROWMOVEMENT
enabled:
ToenableROWMOVEMENTonatablethat’salreadybeencreated,usethis:
IfweweretogobackandredoourearlierseriesofstatementswithROW
MOVEMENTenabledonthetable,ourFLASHBACKTABLEstatementwould
workperfectly,andwe’drestoreourtabletoitsoriginalstate.Onceit’srestored,
wecouldquerythetableandseeallthedatainthattableasitexistedatthetime
wespecifiedinourFLASHBACKTABLEstatement.
Datarestorationis“permanent.”ItinvokesanimplicitCOMMITsothatthe
restoreddataiscommitted.NotethatROWMOVEMENTisn'trequiredwith
flashbackTOBEFOREDROP.
Limitations
YoucannotusetheFLASHBACKTABLEstatementtorestoreolderdatatoan
existingtableifthetablehasbeenstructurallyalteredwiththeALTERTABLE
statementinsuchawaythatitcan’tacceptthefulldefinitionofolderdata.For
example,ifacolumnhasbeendroppedoracolumn’sdatatypechanged,the
FLASHBACKTABLEstatementwon’tsuccessfullyrestoretheolderdata.
MarkingTime
Thereareseveralwaystoidentifythepointatwhichyouwanttorestoredatain
thedatabase.Yousawthreeintheprevioussection,andI’lldiscussthemabit
morehere.
SCN
Thesystemchangenumberisanumericstampthatthedatabaseautomatically
incrementsforeverycommittedtransactionthatoccursinthedatabase.This
includesbothexplicitandimplicitcommits,forallexternalorinternal
transactions.TheSCNisautomaticallymanagedbythedatabaseinrealtime.
EverycommittedtransactionisassignedanSCN.
IfyouwanttodeterminethecurrentSCNatanygivenmomentinthe
database,usethefunction
DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBER.Here’san
example:
Thisexampleshowstherequestandtheanswer.Inthiscase,theSCNis
5896167.Ifyouweretohesitateafewsecondsandrunthesamestatementin
line1again,youmightgetadifferentvaluereturnedforSCN.
TheSCNcanalsobefoundwiththequerySELECTCURRENT_SCNFROM
V$DATABASE,whichisaqueryofthedatadictionary.We’lldiscussthedata
dictionaryinChapter12,butfornow,notethatOracleofficiallyrecommends
thatifyourgoalistoobtainanSCNfromwithinanapplicationorany
comparablecode,thenyoushouldusetheDBMS_FLASHBACK
.GET_SYSTEM_CHANGE_NUMBERfunctionthatwedemonstratedearlier.
Thisrecommendationimpliesthatyoushouldsteerawayfromtappingthe
V$DATABASEviewfortheSCNnumber.The
DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBERfunctionisa
functionwritteninthePL/SQLlanguage,anditispartofthe
DBMS_FLASHBACKpackage.SeeOracle’sreferencemanualonPL/SQL
packagesifyouwanttolearnmore—butyoudon’tneedtodothatforthe
exam.
Eachtimeatransactioniscommittedinthedatabase,theSCNisincremented
andstoredwitheachrowineachtable.
TheSCNforagivenrowcanbefoundinthepseudocolumn
ORA_ROWSCN.Asisthecasewithanypseudocolumn,itcanbeincludedas
anexpressioninanySELECTstatement.Forexample,thefollowing:
returnseachrowinthetableHOUDINI,alongwiththevaluesintheVOILA
column,aswellastheassignedSCNnumberforeachrow.
Timestamp
ATIMESTAMPvaluespecifiesapointintime.TheTIMESTAMPdatatype
storestheyear,month,day,hour,minute,second,andfractionalseconds.A
literalvaluemaybeconvertedtotheTIMESTAMPdatatypewiththe
TO_TIMESTAMPfunction,likethis:
ThisexampleshowstheTO_TIMESTAMPfunction,aliteralvalue,anda
formatmaskthatdefinesthelocationwithintheliteralvalueofeachcomponent
thatformsavalidTIMESTAMPvalue.Notetheuseoffractionalseconds,where
morethantwodigitsareaccepted(uptosix).AlsorecallthatMIistheformat
maskforminutes—notMM,whichistheformatmaskformonths.
TheFLASHBACK_TABLEfunctioncanuseaTIMESTAMPvalueto
specifyapointintimeinthedatabasetowithinthreesecondsofaccuracy.
Ifaspecificpointinthedatabaseisneeded,don’tuseTIMESTAMP—use
SCNinstead.
Ifyouattempttoflashbacktoapointintimeatwhichthedatabasedidnot
exist,you’llgetanerrorindicating“invalidtimestampspecified.”
Ifthetimeyou’rereferencingcloselyalignswithatimeatwhichtheobjector
datadidn’texist,andtheSCN/timestampcorrelationmissesyourtargetand
overshootsintoatimeframeinwhichtheobjectordatadidn’texist,youmay
getanOracleerror.
Youcanuseacombinationofoneormoreconversionfunctionstoaddress
thisdiscrepancy,aswediscussinthenextsection.
ConversionFunctions
SCNnumberscanbeconvertedintotheirequivalentTIMESTAMPvalues,and
viceversa.Theconversionsarenotexact,however,becausetheSCNand
TIMESTAMPvaluesdonotrepresentmomentsintimethatareprecisely
identical.
Theconversionfunctionsare
SCN_TO_TIMESTAMP(s1)TakesanSCNexpressionasinput;
returnsaTIMESTAMPvalueroughlycorrespondingtowhentheSCN
wasset
TIMESTAMP_TO_SCN(t)TakesaTIMESTAMPexpression
representingavalidpastorpresenttimestampasinput;returnsanSCN
valueroughlycorrespondingtowhentheTIMESTAMPoccurred
Forexample,let’sperformtwoseparateconversionsoftimestampvaluesto
SCNvalues:
Inthisexample,weconvertthecurrenttimestampvalue,asdefinedby
SYSTIMESTAMP,toitsSCNequivalent.Wealsoconvertadateinthepastto
itsSCNequivalent.Thefirstcolumn’saliasisNOW;thesecondcolumn’salias
isNOT_NOW.
Notethatanytimereferencedmustbewithinarangethatisrelevanttothe
database.Inotherwords,thefunctionrecognizestimesthatapplytowhenever
thedatabaseinstallationhasbeeninexistence.
Thereisnotadirectone-to-onerelationshipbetweentimestampsandSCN
values.Forexample,supposeyoutakeavalidSCNvalueandconvertit.
Notewhathappenshere:oneSCNgoesin,andaslightlydifferentSCNis
ultimatelyreturned.
Foranyworkeffortinwhichyourequireprecisioninspecifyingtiming,
OraclerecommendsusingSCNnumbersandobtainingthemwiththe
DBMS_FLASHBACKpackagedfunction
GET_SYSTEM_CHANGE_NUMBER,whichwereviewedearlier.Alsonote
thatifyouwanttouseFLASHBACKtorestoretoaspecificpoint,itis
importantthatyouobtainthatpointprecisely.Onewaytodothatisthe
RESTOREPOINTobject,whichwediscussnext.
RESTOREPOINT
Arestorepointisanobjectinthedatabaseyoucreatetorepresentagiven
momentinthedatabase.ThatmomentcanbeidentifiedbyaTIMESTAMPvalue
oranSCN.
Here’sanexampleoftheCREATERESTOREPOINTstatement:
Onceexecuted,youcanusethisasarestorepointrepresentingthemomentat
whichtheCREATERESTOREPOINTstatementwasexecuted.Youcanreferto
therestorepointeitherlaterinthecurrentsessionorinalatersession:
ThisstatementrestorestheHOUDINItabletothepointintimethatcorrelates
tothe“balance_acct_01”restorepoint.
WhenyounolongerneedtheRESTOREPOINT,youcandropit:
Youcanfindexistingrestorepointswiththedatadictionaryview
V$RESTORE_POINT.Notethatusersdonot“own”RESTOREPOINTobjects;
theirscopeistheentiredatabase.Theyexistuntiltheyaredroppedorageoutof
thecontrolfile.
NewtoOracle12cistheabilitytocreateamultitenantcloudarchitecturethat
includescontainerdatabasesandpluggabledatabases,butthosetopicsare
beyondthescopeoftheexam.Itisnotpossibletocreatearestorepointfrom
withinapluggabledatabase.Butyoucancreatearestorepointinacontainer
database.Also,rememberthattheexecutionofaDDLstatementincludesan
impliedCOMMIT.CREATERESTOREPOINTisnoexception.ItisDDL,
andthereforeitsexecutionresultsinaCOMMITtothedatabase.
CERTIFICATIONSUMMARY
Tablesstoreallthedatainadatabase.ViewsarenamedSELECTstatements.
Indexesactjustlikeanindexinabook;theyprovideaseparatelystoredsorted
summaryofthecolumndatathatisbeingindexed,alongwithaddressing
informationthatpointsbacktothesourcetable.Sequencesprovideamechanism
tocountoffprimarykeyvalues.
Viewsareobjectsthatnameaqueryandstorethequeryinthedatabase.
Viewsdonotstoredata.Theylookandactliketables.Viewsmaybebuiltwith
simpleorcomplexSELECTstatements.GROUPBY,HAVING,joinqueries,
andmoreareallpossibleinaVIEWobject’sSELECTstatement.
ASELECTstatementcanbeusedtoqueryaviewjustasitwouldbeusedto
queryatable.ButINSERT,UPDATE,andDELETEstatementswillworkonly
withcertainviews,notallofthem—itdependsonwhethertheview’sstructure
allowsappropriateaccesstothetableortablesuponwhichtheviewisbased.
AnyDMLstatementwillworkwithaviewbasedonasingletable,providedthat
therequiredcolumnsareallincludedintheselectlistoftheSELECTstatement
usedtocreatetheview,thattherowsoftheviewarenotaggregaterows,and
thatallofthetable’sconstraintsarehonored.Aviewisnotnecessarilyrequired
toincludecolumnsthatarerequiredbytheview’sunderlyingtable,butifthe
viewomitssuchcolumns,thenanINSERTstatementcannotaddtotheviewa
valuethatisrequiredbytheunderlyingtable;inaddition,theconstraintsonthe
underlyingtablewillrejecttheINSERT,anditwon’tworkforthatview.
AnySELECTstatementusedtocreateaviewmusthaveclearlydefined
columnnames.Complexexpressionsmustbeassignedacolumnalias;thiswill
becomethenameofthecolumnintheresultingviewobject.
Aninlineviewisavariationonasubquery,inwhichaSELECTstatement
replacesatablereferenceinaFROMclauseofanyDMLstatement.Aninline
viewmustbeenclosedinitsownsetofparentheses.Onceincorporatedina
DMLstatement,theinlineviewbehavesjustlikeaviewobjectwouldbehave.
Iftheunderlyingtableismodified,theviewmayrequirerecompilationusing
theALTERVIEW...COMPILEstatement.
Aviewisbuiltwithaquery,andthatquerymayspecifythenamesof
invisiblecolumnsinatable,assumingthattableconsistsofinvisiblecolumns.
Anyinvisiblecolumnsinthequeriedtablethatarespecifiedintheview’squery
willbeaccessibletoDMLstatementsexecutedagainsttheview.
Viewscanbeusedtohidecomplexityinqueries.Theyarebeneficialfor
creatingaseriesofqueries,onebuiltonanother.Bycreatingaqueryasaview
andthenpublishingaccesstothatview,thedevelopercanretaintheabilityto
laterchangethatqueryonwhichtheviewisbuiltsothatthechangeisreflected
intheview.
Sequenceobjectsarecounters.Theydispensenumbersaccordingtothe
parametersestablishedwhenthesequenceisfirstcreated.
SQLautomaticallydefinespseudocolumnsforeachsequence.The
NEXTVALpseudocolumnincrementsthesequencetothenextavailablenumber
asdefinedinthesequence’sparametersdefinedatthetimethesequenceis
created.TheCURRVALpseudocolumnreferstothelastNEXTVALreference
producedinagivensession.Inanysessionwithasequence,theNEXTVAL
pseudocolumnmustbereferencedbeforetheCURRVALpseudocolumncanbe
referenced.Areferencetoasequencepseudocolumnisavalidexpressionand
canbeinvokedinmanyplaceswhereexpressionsareallowed,withcertain
restrictions.
Anindexisanobjectthatstoresasubsetofpresortedinformationfroma
tableandmaybereferencedautomaticallybySQLtospeeduptheprocessingof
agivenSQLstatement,dependingonhowthatSQLstatementisstructuredand
howtheOracleDatabaseoptimizerchoosestobuilditsexecutionplan.Anindex
canbebuiltononeormorecolumnsinatable.Oncecreated,ifaSQL
statement’sWHEREorORDERBYclausereferencestheindexedcolumn(or
setofcolumns),theoptimizermayusetheindexinitsexecutionplantospeed
uptheprocessingofthestatement.Anindexthatisbuiltonmultiplecolumnsis
knownasacompositeindex.AUNIQUEINDEXcanbeusedtoenforcea
uniqueruleonacolumn.Anindexcanberenderedinvisibletopreventthe
optimizerfromincludingtheindexforconsiderationwhenformingthe
executionplan.
Indexobjectssupportapplicationperformancetuning.Indexescanbebuilton
oneormorecolumnsofatable;multicolumnindexobjectsareknownas
compositeindexes.WhenaqueryreferencesindexedcolumnsinitsWHERE
clause,SQLwillconsiderusingtheindexaspartofitsprocessingstrategy,and
theresultsmaybereturnedmorequickly.
Flashbackoperationscanrestoredatatoapreviouspointintime,withsome
limitations.Thepointtowhichyourestoredatacanbeidentifiedusingvaluesof
datatypeTIMESTAMP,SCN,orRESTOREPOINT.Tablesthathavebeen
droppedbutthatarestillintherecyclebincanberecovered.Anytablethat’s
beenpurgedfromtherecyclebinisgoneforgood.Datathat’sbeenchangedin
existingtablescanberestoredtosomepreviouspointintime,providedthatthe
table’sstructurehasn’tbeenchangedinsuchawaythatthetableisnolonger
abletoreceivetherestoreddata,suchasperhapshavinghadacolumndropped.
✓TWO-MINUTEDRILL
DescribeHowSchemaObjectsWork
Tablesstoreallthedatainadatabase.
ViewsarenamedSELECTstatements.
YoucanselectfromaviewwithSELECT.
Youmaybeabletoinsert,update,anddeletefromaview,
dependingontheview’sstructure.
Anindexspeedsqueriesonatablebycreatingapresortedlookup
listforthetable’sindexedcolumns,alongwithanaddresspointerback
totheindexedtable.
Sequenceskeeptrackofnumbercounterstomakethejobofadding
primarykeysandotheruniquevalueseasier.
CreateSimpleandComplexViewswithVisible/InvisibleColumns
AVIEWisaSELECTstatementthatisstoredinthedatabaseand
assignedaname.
ThecolumnsandexpressionsoftheSELECTstatementusedto
createaviewbecomethecolumnsoftheVIEW.
YoucanuseSELECTstatementsonviewsjustasyouwoulda
table.
YoucanuseINSERT,UPDATE,and/orDELETEstatementson
someviews,dependingontheconstraintsoftheview’sunderlyingtable
ortables,aswellasotherissuessuchaswhethertheview’sSELECT
statementincludesaggregatefunctions.
Aninlineviewisasubquerythatreplacesatablereferenceina
FROMclauseofaDMLstatement.
TheVIEWcanbetreatedlikeatable,withsomelimitations.
AVIEWbasedonatablethatissubsequentlyalteredmayrequire
recompilationwiththeALTERVIEW...COMPILEstatement.
AVIEWcanbebasedonatableconsistingofoneormoreinvisible
columns.
AVIEWcanincludeatable’sinvisiblecolumnsinitsstructureby
namingtheinvisiblecolumnsintheVIEW’squery.
Create,Maintain,andUseSequences
ASEQUENCEisanobjectthatdispensesnumbersaccordingtothe
rulesestablishedbythesequence.
ASEQUENCEspecifiesthestartingnumber,theincrement(which
canbepositiveornegative),andanoptionalrangeofvalueswithin
whichthesequencecangeneratevalues.
Thestartingpointcanbewithintherangeoratoneendofthe
range.
SEQUENCESareidealforpopulatingprimarykeyvalues.
TheNEXTVALpseudocolumnofasequencereturnsthenext
availablenumberinthesequenceandmustbethefirstreferencetothe
sequenceinanygivenloginsession.
TheCURRVALpseudocolumncanreturntheexistingvalueas
alreadydefinedbyNEXTVAL;itcanbereferencedinasessiononly
aftertheNEXTVALreferencehasoccurred.
Apseudocolumnreferencetoasequenceisavalidexpressionand
canbereferencedanywherethatavalidexpressionisallowed.
IfavalidsequencereferencetoNEXTVALoccurswithinaDML
statementthatfails,thesequencestilladvances.
CreateandMaintainIndexesIncludingInvisibleIndexesand
MultipleIndexesontheSameColumns
AnINDEXobjectisbuiltononeormorecolumnsinatable.
TheINDEXstoresdatafromitstable’scolumnsonwhichitisbuilt
andpresortsthatdatainordertospeedfuturequeries.
WhentheDMLstatementsINSERT,UPDATE,andDELETEare
executedonanindexedtablesothattheindexeddataischanged,the
indexisautomaticallyupdatedbyOracle,thusaddingtotheworkloadof
theDMLstatements.
INDEXobjectsarebuiltononeormorecolumns.Multiple-column
indexesarecompositeindexes.
YouusethekeywordsVISIBLEandINVISIBLEtospecifythe
visibilityofanINDEX.ThedefaultisVISIBLE.Youcanspecifythe
visibilityofanindexinCREATEINDEXandALTERINDEX
statements.
PerformFlashbackOperations
FLASHBACKTABLEcanbeusedtorestoreatablethathasbeen
dropped.
Ifatablehasbeendropped,itgoesintotherecyclebin.
Youcaninvestigatethecontentsoftherecyclebintodetermine
whatobjectsareavailableforusebyFlashbackoperations.
Onceanobjecthasbeendropped,ifitisalsopurgedwiththe
PURGEstatement,itisnolongerabletoberecoveredwith
FLASHBACKTABLE.
Inadditiontorestoringadroppedtable,theFLASHBACKTABLE
statementcanbeusedtorestoredatawithinthetableasofaparticular
timeinthedatabase.
Timeinthedatabaseismarkedbyeitherasystemchangenumber,a
RESTOREPOINT,oratimestamp.
SELFTEST
Thefollowingquestionswillhelpyoumeasureyourunderstandingofthe
materialpresentedinthischapter.Chooseonecorrectanswerforeachquestion
unlessotherwisedirected.
DescribeHowSchemaObjectsWork
1.Thedatabaseobjectthatstoreslookupinformationtospeedup
queryingintablesis:
A.ROWID
B.INDEX
C.VIEW
D.LOOKUP
2.ASEQUENCEis
A.Partofatable
B.Usedexclusivelytocreateprimarykeyvalues
C.Istheonlywaytopopulateprimarykeyvalues
D.Noneoftheabove.
3.Alldatabasedataisstoredin:
A.TABLES
B.TABLESandVIEWS
C.TABLES,VIEWS,andSEQUENCES
D.Noneoftheabove
CreateSimpleandComplexViewswithVisible/InvisibleColumns
4.WhichofthefollowingSQLstatementscanalwaysbeexecutedon
anyVIEWobject?(Chooseallthatapply.)
A.SELECT
B.INSERT
C.DELETE
D.UPDATE
5.Reviewthefollowingillustration:
NowreviewthefollowingSQLcode:
WhatwillresultfromanattempttoexecutethisSQLcode?
A.Thestatementwillfailbecauseofanerrorinline3.
B.Thestatementwillfailbecauseofanerrorinline6.
C.Thestatementwillfailbecauseofanerrorinline7.
D.Thestatementwillexecute,andtheviewwillbesuccessfully
created.
6.Reviewtheillustrationfromquestion5andthefollowingSQLcode:
Whatcanbesaidofthiscode?
A.Aftertheviewiscreated,avalidSELECTstatementwillwork
onthePROJECTS_ROLLUPview,butanINSERTwillnot.
B.Aftertheviewiscreated,avalidSELECTandvalidINSERT
statementwillworkonthePROJECTS_ROLLUPview.
C.Theattempttocreatetheviewwillfailbecauseyoucannot
createaVIEWwithaSELECTstatementthatusesaGROUPBY
clause.
D.Theattempttocreatetheviewwillfailbecauseyoucannot
createaVIEWwithaSELECTstatementthatisajoin.
7.Reviewtheillustrationfromquestion5andthefollowingSQLcode:
WhatwillresultfromanattempttoexecutethesetwoSQLstatements?
A.TheCREATEstatementwillfailbecauseitomitsthePURPOSE
columnfromthePROJECTStable.
B.TheINSERTstatementwillfailbecauseofanerroronlines9
and10.
C.TheINSERTstatementwillfailbecausethePROJECT_COST
valuebeinginsertedisnotconsistentwiththeWHEREclauseonline
4.
D.TheCREATEandINSERTstatementswillsuccessfullyexecute.
Create,Maintain,andUseSequences
8.WhichofthefollowingkeywordscannotbeusedwiththeCREATE
SEQUENCEstatement?
A.CYCLE
B.MAXVALUE
C.INCREMENT
D.JOIN
9.Reviewthiscode:
NotethatthefirstINSERTstatementisattemptingtoenterastring
literalof'NOTANUMBER'intoacolumndeclaredwithanumericdata
type.Giventhat,whatwillbetheresultoftheseSQLstatements?
A.OnerowaddedtotheSHIPStable,withaSHIP_IDvalueof1.
B.OnerowaddedtotheSHIPStable,withaSHIP_IDvalueof5.
C.TworowsaddedtotheSHIPStable.ThefirstSHIP_IDis1;the
secondis5.
D.TworowsaddedtotheSHIPStable.ThefirstSHIP_IDis
NULL;thesecondis5.
10.Reviewthiscode:
WhatwillresultfromtheseSQLstatements?
A.TheSELECTstatementwillfailbecausethesequencecanbe
referencedonlyinanINSERTstatement.
B.TheSELECTstatementwillfailbecauseyoucannotreference
theCURRVALpseudocolumnofasequenceuntilafteryouhave
referencedNEXTVALforthesequenceinasession.
C.TheSELECTstatementwilldisplayavalueof1.
D.TheSELECTstatementwilldisplayavalueof3.
CreateandMaintainIndexesIncludingInvisibleIndexesand
MultipleIndexesontheSameColumns
11.Choosethebestanswerfromthechoicesbelow.Anindex:
A.Storesallthedatafromallthecolumnsinanygiventableina
separateobjectandsortsthedataforfasterlookups
B.MayimprovetheperformanceofanUPDATEstatementthat
usesaWHEREclause,iftheWHEREclauseperformsanequality
comparisononanindexedcolumninatable
C.RequiresaseparateINSERTstatementeachtimeyouadddatato
atable—onetimetoaddanewrowtothetable,anothertimetoadd
thecorrespondingandnecessarydatarequiredbytheindex
D.OnlybenefitsaSELECTstatementiftheSELECTreturnsdata
thatisindexed
12.ReviewthefollowingseriesofSQLstatements:
Assumingtherearenoobjectsalreadyinexistencenamed
SUPPLIES_01orSUPPLIES_02priortotheexecutionofthepreceding
statements,whatdatabaseobjectswillresultfromthesestatements?
A.AtablecalledSUPPLIES_02andnothingelse
B.AtablecalledSUPPLIES_02andanindexcalledIX_SU_02
C.AtablecalledSUPPLIES_02andtwoindexescalledIX_SU_01
andIX_SU_02
D.Noneoftheabove
13.Aninvisibleindexisanindexononeormorecolumnsinatable:
A.Whereatleastoneofthecolumnsmustbeinvisible
B.Whereallthecolumnsmustbeinvisible
C.AndisupdatedforanyDELETEstatementsperformedonthe
table
D.AndisupdatedforanySELECTstatementsperformedonthe
table
PerformFlashbackOperations
14.ReviewthefollowingSQLcode:
Whichofthefollowingstatementscouldbeaddedasline11andrecover
thedeletedrowsfromthePO_BOXEStable?
A.FLASHBACKTABLEPO_BOXESTOTIMESTAMP
SYSTIMESTAMP—INTERVAL‘000:00:45’DAYTOSECOND;
B.FLASHBACKTABLEPO_BOXESTOSYSTIMESTAMP—
INTERVAL‘000:00:45’DAYTOSECOND;
C.FLASHBACKTABLEPO_BOXESINTERVAL‘000:00:45’
DAYTOSECOND;
D.FLASHBACKTABLEPO_BOXESTOTIMESTAMP
INTERVAL‘000:00:45’DAYTOSECOND;
15.ReviewthefollowingSQLcode:
WhatstatementwillrecoverthePO_BOXEStableafterthesestatements
areexecuted?
A.FLASHBACKTABLEPO_BOXESTOBEFOREDROP;
B.FLASHBACKTABLEPO_BOXESTOTIMESTAMP
SYSTIMESTAMP—INTERVAL‘000:00:03’DAYTOSECOND;
C.FLASHBACKTABLEPO_BOXESTOBEFORECOMMIT;
D.Noneoftheabove—thetablecannotberecovered.
SELFTESTANSWERS
DescribeHowSchemaObjectsWork
1. B.TheINDEXstoresdataforspeedingupquerying.
A,C,andDareincorrect.ROWIDisapseudocolumn,notadatabase
object.AVIEWnamesaSELECTstatement.LOOKUPissomethingmade
up;it’snotareservedwordinOracle.
2. D.Noneoftheabove.ASEQUENCEisacounterthatisusefulfor
populatingprimarykeysbutcanbeusedforanypurposethedeveloper
wishes,ornot.
A,B,andCareincorrect.ASEQUENCEisastandaloneobjectandis
notpartofatable.Itmaybeusedtopopulateprimarykeysoranyother
purposeforwhichthedeveloperchooses;itismerelyanumbercounter.
Primarykeysmaybepopulatedinanyway,providethevaluesentered
conformtotheparticularprimarykey'sspecifiedconstraints,whichare
usuallythatthekeybeuniqueandNOTNULL.
3. A.Alldataisstoredintables.Evendataaboutthetablesyoucreate
isstoredautomaticallybyOracleSQLinasetofsystem-definedand
system-maintainedtables.
B,C,andDareincorrect.
CreateSimpleandComplexViewswithVisible/InvisibleColumns
4. A.TheSELECTstatementcanbeusedagainstanyview.
B,C,andDareincorrect.TherearemanyreasonswhyanygivenVIEW
mayrejectattemptstoexecutetheINSERT,UPDATE,orDELETE
statement.Forexample,iftheVIEWobjectdoesnotcontainsufficient
accesstoitsunderlyingtables,thenitmightnotprovideallthecolumn
accessrequiredtosatisfytheunderlyingconstraintrestrictionsthatmay
existforanyINSERT,UPDATE,orDELETEstatement.Whilemanyviews
allowalloftheseSQLstatementstowork,it’sentirelypossibletocreatea
VIEWthatdoesnot.
5. A.Theerroronline3isthefailuretogivetheexpressionacolumn
alias.TheVIEWdoesnotassignanametothesecondcolumn,sothe
attempttocreatetheviewfails.
B,C,andDareincorrect.Thesearenoterrors.Line6isfine;theuseof
USINGworkswellhere.Line7isfine;theexpressionthatcalculatesthe
resultofPROJECT_COSTtimes2isfine,andexpressionsinWHERE
clauseshavenorestrictionsinaCREATEVIEWstatement.
6. A.Thesyntaxforcreatingtheviewiscorrect,andanyviewcan—at
aminimum—workwithaSELECTstatement.ButanINSERTwillnot
workwiththisviewsinceitconsistsofaggregaterows,asdefinedbythe
GROUPBYclauseintheview’sSELECTstatement.
B,C,andDareincorrect.TheviewwillnotworkwithanINSERT
becauseoftheGROUPBY.Inotherwords,thereisnowaytoaddsingle-
rowvaluesthroughtheviewsincetheview’saccesstothetablesis,by
definition,attheaggregatelevel.Also,theGROUPBYandJOIN
conditionsfortheSELECTarefine.Youcanmostcertainlycreateviews
withthoseclausesintheSELECTstatement—theyjusthavetheeffectof
limitingthecapabilitiesoftheresultingview,aswe’veseen.Suchviews
won’tacceptINSERTorUPDATEstatements.
7. D.TheCREATEandINSERTstatementswillsuccessfullyexecute.
A,B,andCareincorrect.ThePURPOSEcolumnisnotrequiredto
createtheview—thesubqueryinline2throughline4isavalidSELECT
statement.Thecodeonlines9and10specifiesscalarsubqueriesandis
correct.ThelowerPROJECT_COSTvaluewillnotpreventtheINSERT
statementfromworking;however,itwillpreventtherowfromeverbeing
seenthroughtheMAJOR_PROJECTSview.ASELECTstatementthat
attemptstodisplaythisrowinthefuturecoulddosobyqueryingthe
originaltablePROJECTS,butnottheviewMAJOR_PROJECTS,which
onlyseesrowswithaPROJECT_COSTgreaterthan10000butcertainly
allowsthemtobeinsertedthroughtheviewintotheunderlyingtable.
Create,Maintain,andUseSequences
8. D.JOINisusedinaSELECTstatementthatconnectstwoormore
tables.ButitisnotusedinaCREATESEQUENCEstatement,eventhough
itsultimatepurposemaybetosupporttheintegrityofjoins.
A,B,andCareincorrect.CYCLEspecifieswhetherthesequencewill
repeatarangeonceitreachestheendoftherange.MAXVALUEspecifies
oneendoftherange.INCREMENTspecifiesthenumberbywhichthe
sequencewillincrement.
9. B.Therewillbeonerowinthetable.Thereasonisthatthefirst
INSERTwillfailbecauseoftheattempttoenteracharacterstringintoa
numericcolumn.InthefirstfailedINSERTstatement,thePROJ_ID_SEQ#
sequencegeneratorwillbeinvoked,andtheNEXTVALreferencewilluse
upthefirstnumberinthesequence,whichwillbe1.ThesecondINSERT
willsucceedandgrabthesecondnumberinthesequence,whichwillbe5.
A,C,andDareincorrect.
10. B.Sincethesequencewasjustcreated,NEXTVALmustbereferenced
beforeCURRVAL.Thisisalsotrueifyouweretologoffandendthe
sessionandthenlogbackintorestartthesession—thefirstreferencefor
existingsequencesmustbeNEXTVAL.
A,C,andDareincorrect.Youareallowedtoreferencesequence
generatorsinanySQLstatementwhereexpressionsareallowed.If
CURRVALwerereplacedwithNEXTVAL,thecorrectanswertothis
questionwouldhavebeenC,meaningthattheSELECTstatementwould
havedisplayedavalueof1.
CreateandMaintainIndexesIncludingInvisibleIndexesand
MultipleIndexesontheSameColumns
11. B.AnindexcanpotentiallyspeeduptheWHEREclauseofanyDML
statement,includingtheUPDATEstatement.Comparisonsofequalityare
ideal.
A,C,andDareincorrect.Anindexdoesnotstoreallthedatafromall
therowsofanindexedtablebutinsteadstoresonlythatdatathatis
containedwithintheindexedcolumnorcolumns.Itdoesnotrequirea
separateINSERTstatement;instead,anyfutureINSERTswillperform
automaticindexmaintenanceasrequiredwithoutanyextraeffortfromyou.
TheSELECTstatementdoesnotneedtoreturnindexedinformationforthe
indextobeuseful.
12. B.Whileallthestatementswillexecutesuccessfully,thefirstDROP
statementwilldropthetableSUPPLIES_01,whichwillcausetheindex
IX_SU_01tobedroppedaswell.Inotherwords,theDROPTABLE
SUPPLIES_01statementhastheeffectofdroppingthetable
SUPPLIES_01aswellastheindexIX_SU_01.ThetablesSUPPLIES_02
andIX_SU_02willremainattheend.
A,C,andDareincorrect.
13. C.TheindexisupdatedforanyDELETEstatementsperformedonthe
table.Invisibleindexesarestillmaintained,eventhoughtheyareinvisible.
A,B,andDareincorrect.Aninvisibleindexhasnothingtodowith
invisiblecolumnsinthetable.TheuseofaSELECTstatementnever
resultsinamaintenanceactionofanindex,regardlessofthevisibilityof
theindex.
PerformFlashbackOperations
14. A.Thisisthecorrectsyntax—theTOTIMESTAMPclausewiththe
expressionthatstartswiththecurrentdateandtimeandsubtractsan
intervalof45seconds.
B,C,andDareincorrect.OptionBismissingthekeyword
TIMESTAMP.OptionCismissingthekeywordsTOTIMESTAMPand
includesanincompleteexpressionthatonlyrepresentsatimeintervalof45
seconds.OptionDhasthekeywordTIMESTAMPbutalsohasthe
incompleteexpressionthatonlyincludestheintervalvalueandnothing
more.
15. D.Noneoftheabove.ThePURGEstatementonline8preventsany
recoveryfrombeingpossible.PURGEcleansouttherecyclebinofthe
objectsspecifiedinthePURGEstatement,fromwhichFLASHBACK
TABLErecoversobjects.
A,B,andCareincorrect.WereitnotforthePURGE,optionAwould
becorrect.OptionBissyntacticallycorrect,exceptforthePURGE
statementinthecodesample.ThereisnoBEFORECOMMIToptionfor
FLASHBACKTABLE.
11
T
UsingtheSetOperators
CERTIFICATIONOBJECTIVES
11.01DescribeSetOperators
11.02UseaSetOperatortoCombineMultipleQueriesintoaSingleQuery
11.03ControltheOrderofRowsReturned
✓Two-MinuteDrill
Q&ASelfTest
hischapterdescribesthesetoperatorsinSQL.Setoperatorsworkwithsets
ofoutputdatafromtwoormoreSELECTstatements.Theycombine
standaloneSELECTstatementsinwaysthatcannotbedonewithjoinsorother
conventionalmethodsinSQL.
SetoperatorsareidealforavarietyofsituationswhereaSELECTstatement’s
outputcanbecombinedwithotherdatathatisn’tnecessarilyrelatedthrougha
structuredkeyrelationshipbutcanstillbeblendedintoonecompleteoutputdata
set.AseriesofSELECTstatementscombinedwithsetoperatorsmayincludea
singleORDERBYclauseattheendoftheseriesofSELECTstatements.Set
operatorsshouldnotbeconfusedwiththereservedwordSETthatisusedwith
SQLstatementslikeUPDATE.Thesetoperatorshavenothingtodowiththe
keywordSETanddon’tusetheSETvariable_nameclause.
CERTIFICATIONOBJECTIVE11.01
DescribeSetOperators
Therearefoursetoperators:UNION,UNIONALL,INTERSECT,andMINUS.
SetoperatorscombinetwoormoreseparateSELECTstatementssothattheir
outputismergedinsomemanner.Eachsetoperatormergesthedataina
differentway.ThesetoperatorsaredescribedinTable11-1andsummarizedin
Figure11-1.
TABLE11-1 TheSetOperators
FIGURE11-1 Thesetoperatorsinaction
TheUNIONoperatormergestheresultingrowsetsofoneSELECTstatement
withtheresultingrowsetsofanothersothatalltherowsfrombothSELECT
statementsareincludedinthefinaloutput.UNIONalsoeliminatesanyduplicate
recordsthatmightresultinthecombinedoutput.
UNIONALLdoesthesamethingasUNION,exceptitdoesnoteliminate
duplicaterows.
INTERSECTcombinesbothsetsofrowssothatonlythoserowsthatwere
presentinbothSELECTstatementsappearinthefinalresult.
MINUSstartswiththefirstsetofrowsandthenusesthesecondSELECT
statement’srowsettoseewhetheranyduplicatesoccur.Iftheydo,those
duplicatesarecompletelyremoved,leavingonlythoserowsthatuniquelyexist
inthefirstSELECTstatement’srowset.
TheonlyrulesforensuringthattheSELECTstatementswillcombine
successfullywithanyofthesetoperatorsareasfollows:
Thenumberofexpressionsselectedintheselectlistsmustbe
identicalineachSELECTstatement.
ThedatatypesofeachexpressionmustmatchsothateachSELECT
statement’sfirstexpressionsharesthesamedatatypegroupwiththe
otherfirstexpressions,eachsecondexpressionsharesthesamedatatype
groupwiththeothersecondexpressions,andsoon.Bydatatypegroup,
wemeandatatypesthateitherareidenticalorcanbemadetobe
identicalbySQLthroughautomaticdatatypeconversion.
LargedatatypessuchasBLOBandCLOBcannotbeused.
TheORDERBYclausecannotbeincludedintheSELECT
statements—exceptafterthelastSELECTstatement.
YoucancombinecomplexSELECTstatementswiththesetoperators.
ExamplesincludeSELECTstatementsthatinvolvemultitablejoins,
subqueries,aggregatefunctions,and/orGROUPBYclauses.
TheSELECTstatementsarenotrequiredtohaveanysortofPRIMARYKEY
/FOREIGNKEYrelationships.Theirtablesandcolumnsdon’thavetobe
identicalinanyotherway.Forinstance,theydon’thavetobenamedwiththe
samenames—noneofthatapplies.EachindividualSELECTstatementcanbea
complete,standalonestatement(butwithoutanORDERBYclauseunlessitis
afterthefinalSELECTstatementintheset),withallthecomplexitiesofany
SELECTstatement,includingGROUPBYclauses,subqueries,andeverything
thatformsacompleteSELECTstatement.Aslongasthenumbersofcolumns
areidentical,andtherespectivedatatypegroupsmatchup,thesetoperatorswill
performasintended(subjecttotherestrictionswejustidentified).
CERTIFICATIONOBJECTIVE11.02
UseaSetOperatortoCombineMultipleQueriesinto
aSingleQuery
Let’slookatthesyntaxforthesetoperators.Eachissimple—eachsetoperator
connectstwoSELECTstatements,causingthemtobehaveasone.
UNION
TodemonstratetheuseofUNION,we’lllookatanexerciseinwhichweare
tryingtocombinee-mailaddressesfromdifferenttables.First,we’lllookatone
ofthetwotablesshowninFigure11-2,theCONTACT_EMAILStable.
FIGURE11-2 DiagramoftheCRUISE_CUSTOMERSand
CONTACT_EMAILStables
NoticethecolumnEMAIL_ADDRESSintheCONTACT_EMAILStable.
Let’sgetadatalisting:
Next,let’slookatanothertablecalledONLINE_SUBSCRIBERS,shownin
Figure11-3.YoucanseethatithasacolumncalledsimplyEMAIL.Let’sgeta
datalistingforthattableaswell:
FIGURE11-3 DiagramoftheONLINE_SUBSCRIBERStable
TodemonstratethesetoperatorUNION,we’llcreatetwoSELECT
statementswithsimilarselectlists—butbysimilar,wemeanonlythattheselect
listsareidenticalinthenumberofexpressionsinthelistandtheirrespective
datatypegroups.ThefirstSELECTwillincludeaWHEREclausetolimitour
rowsto'Valid'data.Let’sgiveitatry(linenumbersadded):
Noticehowwe’vestructuredthisUNIONquery:
BothSELECTstatementshavetwoexpressionsintheirselectlists
(line1andline5).
Thedatatypesofeachlist’sfirstexpressionarethesame:
CONTACT_EMAIL_ID(line1)andONLINE_SUBSCRIBER_ID(line
5).Inthiscase,theyarebothnumeric.
Thedatatypesofeachlist’ssecondexpressionarethesame:
EMAIL_ADDRESS(line1)andEMAIL(line5).Inthiscase,theyare
bothcharacterstrings.
Here’stheoutputfromourUNION:
NoticethefirstcolumnofoutputundertheheadingCONTACT_EMAIL_ID.
TheoutputshownincludesdatafrombothSELECTstatements,whichmeans
you’reseeingvaluesfromthecolumnCONTACT_EMAIL_IDofthefirst
SELECTstatementandfromthecolumnONLINE_SUBSCRIBER_IDofthe
secondSELECTstatement.
Logically,though,there’saproblem:thevaluesforCONTACT_EMAIL_ID
andONLINE_SUBSCRIBER_IDdon’treallyrepresentthesameinformation.
Bothareprimarykeyvaluesbutfordifferenttables,andtheydon’treallybelong
inthesamecolumn—theyrepresenttotallydifferentvalues.Theevidenceofthis
problemispresentinthelistofe-mailaddresses—weonlywantedasinglelist
ofuniquevalues,butinsteadwehaveaduplicationofatleastonee-mail
address.ThereasonisthatUNIONlooksattheentirerowofoutputfromeach
SELECTandshowsuniqueoccurrencesofthecombinedsetofcolumnsinthe
row.Thefirstoccurrenceofthee-mailaddress“watcher@foursigma.org”hasa
firstcolumnvalueof2,andthesecondhasafirstcolumnvalueof3,soasfaras
theUNIONisconcerned,theseareuniquerowsofdata.
Let’sremovetheillogicalreferencestothefirstcolumninthisexampleand
producesomethingthatmakesmoresense:
Here’stheresult:
TheresultofUNION(line4)isacombinationoftheoriginalrows,withany
duplicatevaluesremoved.Inthiscase,thereweretworowsforthee-mail
address“watcher@foursigma.org,”butonlyoneisshowninourfinalresult.
NoticethattheoutputcolumnheadingsdisplaythefirstSELECTstatement’s
expressions(line1).We’llhavemoretosayaboutthatwhenwediscussORDER
BYandcolumnreferences.Fornow,let’smoveontotheUNIONALLset
operator.
UNIONALL
TheonlydifferencebetweenUNIONandUNIONALListhatduplicatevalues
aren’tremovedfromUNIONALL.IfweweretoexecutethesameSQL
statementfromthepreviousexamplewiththeUNIONALLsetoperator,it
wouldlooklikethis:
Herearetheresults:
Noticethattheduplicateentryisincluded.Thevaluefor
"watcher@foursigma.org"appearsinbothtables,soitappearstwiceinthe
outputofourUNIONALLuseoftheSELECTstatement.
Noticealsothattheorderingoftherowsistotallydifferent—rememberthat
withoutanexplicitORDERBYclause,youcanneverguaranteetheorderingof
rowsinanySELECTstatement.We’lllookathowtouseORDERBYlaterin
thischapter.First—let’slookatanothersetoperator.
INTERSECT
ThesetoperatorINTERSECTlooksforcommonvaluesamongtherowsofthe
SELECTstatement.Let’schangeourexampletouseINTERSECT:
Herearetheresults:
JustonerowwascommonbetweenthetwoSELECTstatements.
INTERSECTwilleliminateduplicaterows.IfoneorbothSELECTstatement
rowsetscontainduplicateswithinitsownsetofrows,theresultingoutputfrom
INTERSECTwilleliminatethoseduplicates.
MINUS
ThefinalsetoperatorisMINUS.Uptonow,theresultsofthesetoperators
wouldbethesameregardlessofwhichSELECTstatementwasplacedfirst
beforethesetoperator.That’snotthecasewithMINUS,however.MINUSwill
startwiththefirstSELECTstatementandremoveanyrowsfromthatSELECT’s
outputthatmighthappentoappearinthesecondSELECT’soutput.Theresults
maydiffer,dependingonwhichSELECTisplacedfirstandwhichisplaced
second.
Usingoursameexample,hereistheSELECTfromCONTACT_EMAILS
first:
Herearetheresults:
Butnowlet’sreversetheplacementoftheSELECTstatements:
Herearetheresults:
Noticethattheresultsarecompletelydifferent.Noticealsothatthecolumn
headingisdifferent—thecolumnhastakentheheadingofthefirstSELECT
statement,asitalwaysdoes,butnowthefirstSELECTstatementisdifferent,
andalongwithit(inthisexample)soistheheading.
Combinations
Thesetoperatorsmaybeusedinmultiplecombinations,suchas
Suchcombinationscanbecontinuedindefinitely.
Setoperatorshaveequalprecedenceamongthemselves,meaningthatthey
willallexecutefromstarttofinishintheorderthattheyappearintheSELECT
statement.Tochangetheorderofexecution,useparentheses,likethis:
Justbesuretousetheparenthesessothat:
Thecodeenclosedisastandalonequeryanddoesnotincludean
ORDERBYclause.
Theenclosedcodeisplacedintotheouterqueryasthoughitwerea
single,standaloneSELECTstatement,withouttheORDERBYclause.
IfanORDERBYisdesired,itmustbethefinalclauseintheentire
seriesofstatements.
Followthoserules,andyoucanconnectasmanySELECTstatements
togetherasrequired.
Thesetoperatorsareusefulforrequirementstocombinethedataoftwotables
intooneoutputresultsetwhenthetwotableshavenoprimarykey/foreignkey
relationshipwitheachother.
CERTIFICATIONOBJECTIVE11.03
ControltheOrderofRowsReturned
AswithanySELECTstatement,theORDERBYclauseistheonlywayto
determinetheorderingofrowsthatappearinoutput.Aswealreadyknow,the
ORDERBYclausedetermineshowtosortrowsbyidentifyingaseriesofoneor
moreexpressionsthathavesomethingtodowiththetable—ortables—thatare
involvedwiththeSELECTstatement.Oftentheexpressionissimplyacolumn
withineachrow,butORDERBYalsoacceptscomplexexpressions.These
expressionsmayinvolveoneormorecolumnsandperformsomesortof
transformationonthedata.
However,whensetoperatorsareinvolved,there’sabitofanissue.Howdo
youidentifydataintherowswhentherearemultipletablesusingdifferent
columnnames?
TheansweristhattheORDERBYclauseisabitrestrictedinthissituation.
Theclauseisrestrictedtoidentifyingcommonexpressionitemsintheselectlist,
andnothingmore.Therearetwowaystoidentifythem,andwe’veseenthese
earlier—byreferenceandbyposition.Thefollowingsectionsshowexamplesof
theseapproacheswiththesetoperators.
ORDERBY—ByPosition
OnewaytosortrowsofoutputthatresultfromaseriesofSELECTstatements
combinedwithsetoperatorsistousethe“byposition”approach.Here’san
example:
Theprecedingexamplecombinesrowsfromtwotables.Thefirstqueryhas
twoexpressions:
Astringliteral,'Individual'
Aconcatenationoftwocolumns,LAST_NAMEand
FIRST_NAME,separatedbyacommaandaspace
ThesecondSELECTstatementhastwoexpressions:
ThecolumnCATEGORY
ThecolumnVENDOR_NAME
Herearetheresults:
WecansorttherowswithanORDERBYthatidentifiesthepositionwithin
theselectlistoftheexpressionwewanttosortby,likethis:
Hereistheresult:
Remember,whenusingORDERBYwithaseriesofSELECTstatements
connectedwithsetoperators,youcanuseORDERBYonlyonce,attheend.
ORDERBY—ByReference
ThereisanotherwaytouseORDERBYwithsetoperators.ORDERBY
referenceiswhenyounameoneofthecolumnsintheSELECTstatement’s
expressionlist.Whenusingsetoperators,thecolumnnamesusedinthefirst
SELECTstatementareinforce.Usingourearlierexample,let’saddcolumn
aliasestoourfirstSELECTstatement,andwe’llbeabletouseORDERBY:
NotethecolumnaliasPOINT_OF_CONTACTthatisspecifiedattheendofline
2andusedintheORDERBYinline8.Herearetheresults:
So,eitherthe“byposition”or“byreference”approachworkswithORDER
BY.JustbesureyoumakeitthelastclauseoftheentireseriesofSELECT
statements.
IfyoucombineaseriesoftwoormoreSELECTstatementswithset
operators,yourORDERBYclausemustbethefinalclauseandcan
specifycolumnsbynameonlyifitusesthecolumnnamesfromthefirst
SELECTstatement,regardlessofhowmanySELECTstatementsmight
beconnectedwithsetoperators.
CERTIFICATIONSUMMARY
EachsetoperatorcombinesrowsfromtwoindependentSELECTstatements.Set
operatorsallowyoutocombinerowswithoutajoinbymergingentiresetsof
rowsbasedsolelyonensuringthatthenumberofexpressionsanddatatypes
involvedmatchup.
ThesetoperatorsincludeUNION,UNIONALL,INTERSECT,andMINUS.
EachdiffersinhowitreturnsrowsthatappearinoneSELECTstatementthat
mightalsoappearintheotherSELECTstatement.
UNIONcombinesrowsfromtheSELECTstatementsandpresentsone
uniquesetofrowsfromthecombinedSELECTstatements.Iteliminates
duplicaterowsthatmightappearasaresultofthecombination.UNIONALL
combinesbothsetsofrowsandincludesanyduplicatesthatmightoccurasa
resultofthecombinationoftherowsets;thatis,itdoesnoteliminateany
duplicaterows.INTERSECTlooksonlyforrowsthatarepresentinbothsetsof
rowsreturnedbytheSELECTstatements.Onlytheduplicatesbecomethe
outputfromtheSELECTstatement.MINUStakesthesetofrowsfromthefirst
SELECTandremovesanyforwhichduplicatesexistinthesecondset.
YoucancombineseveralSELECTstatementswithasmanysetoperatorsas
youwant.Theywillexecuteoneaftertheother,unlessyouchoosetooverride
thatbehaviorusingparentheses.
EachSELECTstatementcanbeacomplexquery,withmultiplejoins,
subqueries,andGROUPBYclauses.However,onlyoneORDERBYclauseis
allowed,anditmustbeattheendoftheseriesofSELECTstatementsandset
operators.
ORDERBYwithsetoperatorscansortrowsbypositionorreference.Ifby
reference,thefirstSELECTstatement’sexpressionnamesareineffectforthe
entireseriesofSELECTstatements.Youcanusecolumnaliasesinthefirst(or
any)SELECTstatementifyouwant,butitisnotrequired.
✓TWO-MINUTEDRILL
DescribeSetOperators
UNIONcombinestheoutputoftwoSELECTstatements,
eliminatinganyduplicaterowsthatmightexist.
INTERSECTcombinestheoutputoftwoSELECTstatements,
showingonlytheuniqueoccurrencesofdatapresentinbothrowsets
andignoringanythingthatdoesn’tappearinbothsets.
MINUStakesthefirstSELECTstatement’soutputandsubtracts
anyoccurrencesofidenticalrowsthatmightexistwithinthesecond
SELECTstatement’soutput.
UNIONALLdoesthesamethingasUNIONbutdoesnoteliminate
duplicaterows.
UseaSetOperatortoCombineMultipleQueriesintoaSingle
Query
ThesetoperatorsareplacedbetweentwoSELECTstatements.
ThetwoSELECTstatementscanbesimpleorcomplexandcan
includetheirownGROUPBYclauses,WHEREclauses,subqueries,and
more.
TheORDERBYclause,ifused,mustbethefinalclauseofthe
combinedSELECTstatements.
YoucanconnectmultipleSELECTstatementswithmultipleset
operators.
Thesetoperatorshaveequalprecedence.
Youcanuseparenthesestooverridesetoperatorprecedence.
ControltheOrderofRowsReturned
IfanORDERBYclauseisused,itmustbeplacedattheendofthe
SQLstatements.
MultipleSELECTsthatareconnectedwithsetoperatorsmaybe
sortedbypositionorreference.
WhenusingORDERBYreference,thecolumnnameinforceis
whatevercolumnnameexistsinthefirstSELECTstatement.
SELFTEST
Thefollowingquestionswillhelpyoumeasureyourunderstandingofthe
materialpresentedinthischapter.Chooseonecorrectanswerforeachquestion
unlessotherwisedirected.
DescribeSetOperators
1.Thesetoperatorsdonotincludewhichoneofthefollowingkeywords?
A.ALL
B.SET
C.MINUS
D.UNION
2.Youaretaskedwithcleaningupadatabaseapplication.Therearetwo
tablesinthedatabase:ORDERScontainscompletedORDERS,and
ORDER_RETURNScontainsduplicateinformationforallORDERSthat
werelaterreturned.Yourgoalistofindoutwhetheranyrowsin
ORDER_RETURNSexistthatwereneverintheORDERStabletobegin
with.Whichofthefollowingsetoperatorsshouldyouuse?
A.ALL
B.SET
C.MINUS
D.UNION
3.Reviewthefollowingillustrations:
Next,reviewthefollowingSQLcode:
WhatwillresultfromanattempttoexecutethisSQLstatement?
A.ItwillfailwithasyntaxerrorbecauseoftheTO_CHAR
conversionfunctiononline1.
B.Itwillfailbecauseofthetablealiasinlines1and2,which
cannotbeusedinthiscontext.
C.Itwillfailwithasyntaxerroronline3becauseyoucannotuse
anORDERBYinthiscontext.
D.Itwillexecutesuccessfully.
4.WhencombiningtwoSELECTstatements,whichofthefollowingset
operatorswillproduceadifferentresult,dependingonwhichSELECT
statementprecedesorfollowstheoperator?
A.MINUS
B.UNIONALL
C.INTERSECT
D.UNION
5.Whichofthefollowingstatementsaboutsetoperatorsistrue?Choose
thebestanswer.
A.IfyouaddthereservedwordALLtotheendofanysetoperator,
itwillchangethebehaviorofthesetoperatorbyremovingduplicate
rows.
B.SetoperatorscanbeusedtocombineINSERTstatements.
C.YoucanconnecttwoSELECTstatementswithonesetoperator.
D.TheUNIONsetoperatorhasprecedenceovertheothers.
UseaSetOperatortoCombineMultipleQueriesintoaSingle
Query
6.Reviewthefirsttwoillustrationsfromquestion3,andthenreviewthis
SQLcode
Howmanyrowswillresultfromthisquery?
A.0
B.1
C.3
D.6
7.Reviewthefirsttwoillustrationsfromquestion3,andthenreviewthis
SQLcode:
Howmanyrowswillresultfromthisquery?
A.0
B.4
C.6
D.Itwillnotexecutebecauseitwillfailwithasyntaxerror.
8.Reviewthefirsttwoillustrationsfromquestion3,andthenreviewthis
SQLcode:
Howmanyrowswillresultfromthiscode?
A.1
B.2
C.4
D.6
9.Reviewthefirsttwoillustrationsfromquestion3,aswellasthe
ONLINE_SUBSCRIBERStableinFigure11-3,andthenreviewthisSQL
code:
WhatwillhappenwhenthisSQLstatementisexecuted?
A.Itwillfailwithasyntaxerrorbecauseyoucannotusean
aggregatefunctionlikeCOUNT(*)inline1inthiscontext.
B.Itwillfailwithasyntaxerrorstartingatline4.
C.Itwillexecute,butitwillnotperformasintendedbecausethe
secondSELECTstatementwithinthesubqueryonline6willnot
execute;onlythefirstSELECTinthesubqueryonline4willexecute.
D.Itwillexecutesuccessfully.
10.Reviewthefirsttwoillustrationsfromquestion3,aswellasthe
ONLINE_SUBSCRIBERStableinFigure11-3,andthenreviewthisSQL
code:
WhatwillhappenwhenthisSQLstatementisexecuted?
A.Itwillfailwithanexecutionerroronline1.
B.Itwillexecute,buttheUNIONwillnotworkasexpected.
C.Itwillexecuteanddisplayonecolumnunderthe“DateAdded”
heading.
D.Itwillexecuteanddisplayonecolumnunderthe
“LAST_ORDER”heading.
11.Reviewthefirsttwoillustrationsfromquestion3,aswellasthe
ONLINE_SUBSCRIBERStableinFigure11-3,andthenreviewthisSQL
code:
WhatwillhappenwhenthisSQLstatementisexecuted?
A.Itwillfailwithageneralsyntaxerror.
B.Itwillfailwithanexecutionerror.
C.Itwillexecute,buttheINTERSECTwillnotworkcorrectly.
D.Itwillexecuteandrepeatthevalue'Towel'foreachrowofthe
ONLINE_SUBSCRIBERStable.
12.Reviewthefirsttwoillustrationsfromquestion3,aswellasthe
ONLINE_SUBSCRIBERStableinFigure11-3,andthenreviewthisSQL
code:
WhichofthefollowingaretrueaboutthisSQLstatement?(Choose
two.)
A.TheGROUPBYclauseonline7isnotallowedhere.
B.TheB.LAST_ORDERreferenceattheendofline6refersto
dataincludedintheADDEDcolumnreferredtoinline5.
C.TheJOINattheendofline2isnotallowedinthiscontext.
D.Thestatementissyntacticallycorrectandwillexecute
successfully.
ControltheOrderofRowsReturned
13.Reviewthefirsttwoillustrationsfromquestion3,aswellasthe
ONLINE_SUBSCRIBERStableinFigure11-3,andthenreviewthisSQL
code:
WherecanyouaddanORDERBYtothiscode?(Choosetwo.)
A.Attheendofline5beforetherightparenthesis
B.Betweenlines5and6
C.Afterline7
D.Nowhere
14.TheORDERBYclausecanbeincludedinaSELECTwithsetoperatorsif:
A.ItfollowsthefirstSELECTstatement.
B.ItfollowsthefinalSELECTstatement.
C.ItisusedineachSELECTstatementanditsORDERBY
expressionsmatchindatatype.
D.TheORDERBYclausecannotbeusedinaSELECTwithset
operators.
15.Reviewthefirsttwoillustrationsfromquestion3;thenreviewthisSQL
code:
WhichofthefollowingarevalidORDERBYclausesforthisquery?
(Choosetwo.)
A.ORDERBYAISLE
B.ORDERBY"LastOrder"
C.ORDERBYSECTION
D.ORDERBY1
SELFTESTANSWERS
DescribeSetOperators
1. B.ThekeywordSETisnotusedwiththesetoperators.
A,C,andDareincorrect.ALLispartoftheUNIONALLclause.
MINUSandUNIONarebothsetoperators.
2. C.MINUSiswhatyouwoulduse.Thatisthesetoperatorwith
whichyoucanremoverowsfromonetablethatarealsopresentinthe
secondtable,resultinginoutputthatshowsrowsfromthefirsttablethatare
notpresentinthesecond.
A,B,andDareincorrect.ALLisnotafullsetoperator;itworkswith
UNIONALLbutisnotasetoperatoronitsown.SETisnotasetoperator.
UNIONwouldcombinerecordsfrombothtables,whichisnotwhatis
desiredhere.
3. C.TheORDERBYofthefirstSELECTstatementinline3is
incorrectandcausesthestatementtofail.
A,B,andDareincorrect.TheTO_CHARconversionfunctioninline1
iscorrectsyntax.ItensuresthatthedatatypesforLAST_ORDERand
ADDEDcorrespondtoeachother.Thetablealiasonlines1and2isfine.
Buttheentirestatementwillnotexecuteforthereasonexplainedforthe
correctanswer.
4. A.Theonlysetoperatorthatchangesitsendresultbasedonwhich
SELECTstatementprecedesorfollowsthesetoperatorisMINUS.
B,C,andDareincorrect.
5. C.YoucanconnecttwoSELECTstatementswithonesetoperator.
A,B,andDareincorrect.ThereservedwordALLworksonlywith
UNION.SetoperatorscancombineonlySELECTstatements,notother
SQLstatements.Allsetoperatorshaveequalprecedence;onlyparentheses
canbeusedtooverridesetoperatorprecedence.
UseaSetOperatortoCombineMultipleQueriesintoaSingle
Query
6. A.Norowswillresult.Thereasonisthatwe’retryingtointersect
rows,whichmeanstoshowonlythoserowsthatarecommonbetweenthe
tworowsets.Whilebothtablesshareavalueof'Towel',theSELECT
statementsareincludingtheNUMandCAT#columnsfromthetwotables.
InthetworowsforTowel,thevaluesforNUMandCAT#don’tmatch.The
resultisthatneitherrowthatincludes'Towel'isacompletematch.
B,C,andDareincorrect.
7. B.Thefirstselectwillproduceonerow.Thesecondwillproduce
threerows.TheUNIONALLsetoperatorwillcombinetheresultsand
returnfourrows.
A,C,andDareincorrect.Thesyntaxisfineandthestatementwill
execute.
8. A.Onlyonerowwillresult,asexplainednext.
B,C,andDareincorrect.ItmightbetemptingtohavechosenoptionB,
sincethefirstandsecondSELECTstatementcombinationswiththe
UNIONALLwillproducetworowscontainingthevalue'Towel',andso
willthesecondUNIONALL.ButtheINTERSECTwilleliminateduplicate
rowsandreturnonerowfor'Towel'.
9. D.Itwillexecutesuccessfully.Setoperatorsareperfectlyacceptable
inasubquery.
A,B,andCareincorrect.
10. A.SinceweknowfromthedatalistingsthattheresultsoftheSELECT
statementswiththesetoperatorUNIONwillproducemultiplerows,the
statementwillfail,sinceascalarsubqueryiswhatisexpectedherebythe
outerquery.
B,C,andDareincorrect.UNIONisfine,buttheendresultcauseda
problemforreasonsunrelatedtotheUNION.Thecolumnheadingissues
don’tapply,butifthesubqueryhadnotproducedtheexecutionerror,
optionsCandDwouldstillbeincorrect—theheadingwouldbea
concatenatedversionoftheentirestringofcharactersformingthe
subquery.
11. D.Wecantellfromthedatalistingthatthesubquerywillreturnonevalue
representingtheINTERSECTofbothqueries.That,andthefactthatthe
subquerywillreturnjustonecolumninthatonerow,makesthisascalar
subquery,albeitariskyone,sincethere’snoguaranteethatitwillalways
executeasascalarsubquery.Butitwillworkgiventhedatalistings,and
thesubquerywillperformasthoughitwerealiteralvaluewithintheouter
SELECT,returningthesameresultforeachrowofthe
ONLINE_SUBSCRIBERStable.
A,B,andCareincorrect.Thereisnothingwrongsyntacticallywiththe
SQLstatement,andaswediscussindescribingthecorrectchoice,because
ofthedatalistingsweareprovidedwith,wecantellthattherewillbeno
executionerrorseither.TheINTERSECTwillperformjustfine.
12. BandD.ThisisavalidSQLstatementthatwillexecutesuccessfully.The
subqueryonlines3through5isthecompleteUNIONandistreatedasan
inlineviewinthiscontext.Assuch,itbehaveslikeanyotherinlineview
andisperfectlyfineinthiscontext.Attheendofline5isatablealiasB
thatisgiventotheinlineview,andthattablealiasisusedinline6to
identifythecolumnoftheinlineviewcalledLAST_ORDER,which
representsthefirstcolumnofthecombinedSELECTstatements,including
theADDEDcolumn.
AandCareincorrect.GROUPBYisallowed,asistheJOIN,for
reasonsexplainedunderthecorrectchoice.
ControltheOrderofRowsReturned
13. AandC.TheORDERBYcangoattheendoftheinlinevieworatthe
endoftheentireSQLstatement.
BandDareincorrect.
14. B.TheORDERBYisoptional,butifused,itmustbethelastclauseinthe
entireseriesofSELECTstatements.
A,C,andDareincorrect.ORDERBYcannotbeusedinanySELECT
statementswithinaseriesofSELECTstatementsconnectedbyset
operators.Itcanbeplacedattheendonly,followingthefinalSELECT
statement.
15. CandD.AnyORDERBYthatusesthe“byreference”techniquemust
referencecolumnnamesofthefirstSELECTstatement.So,ORDERBY
SECTIONisvalid.Also,the“byposition”isaccepted,soORDERBY1is
good.
AandBareincorrect.TheAISLEcolumnnameisn’trecognized,since
itisn’tacolumninthefirstSELECTstatement.Thesameistrueforthe
“LastOrder”columnalias.
12
T
ManagingObjectswithDataDictionary
Views
CERTIFICATIONOBJECTIVES
12.01QueryVariousDataDictionaryViews
✓Two-MinuteDrill
Q&ASelfTest
hischapterdescribesavaluabletoolthatallOracleprofessionalsshould
understand.ThedatadictionaryisOracle’sbuilt-inreal-timereferencesource
forallinformationabouttheapplicationsthatyoubuildinyourdatabase.Armed
withthefullcapabilitiesofthedatadictionary,youcanobtaininformationonthe
databaseobjectsthathavebeencreatedwithinyourdatabase,whocreatedthem,
when,andmuchmore.
CERTIFICATIONOBJECTIVE12.01
QueryVariousDataDictionaryViews
Thedatadictionaryisacollectionofdatabasetablesandviews.Itis
automaticallybuiltandpopulatedbytheOracledatabase.Theinformationstored
inthedatadictionaryincludesthefulldescriptionofallthedatabaseobjectsyou
createaspartofyourapplication,includingtables,views,indexes,constraints,
sequences,andmore.Inotherwords,theresultofeachDataDefinition
Language(DDL)statementyou’vestudiedinthisbookisrecordedinthe
dictionary,andtheinformationisautomaticallymaintainedbytheOraclesystem
inrealtimeasyouchangedatabaseobjectsandtheirstructures.
Theinformationstoredinthedatadictionaryincludes(butisnotlimitedto):
Thenamesofdatabaseobjects,theirowners,andwhentheywere
created
Thenamesofeachtable’scolumns,alongwithdatatypes,precision,
andscale
Anyconstraints
Views,indexes,andsequences
Thischapterwillexplorehowthedatadictionaryisstructuredandwhatsort
ofinformationiscontainedwithinit.We’llgothroughsomeexamplesofhowto
extractvaluableinformationandusethedatadictionarytoassistyouasyou
buildyourapplications.
Thedatadictionaryisoftenreferredtoasmetadata,whichmeans“dataabout
data.”Thatiswhatthedatadictionaryis—it’sacomprehensivedetaileddatabase
thattrackseverythingthereistoknowaboutthedatabaseapplicationsyoucreate
withintheOraclesystem.Everytimeyoucreateadatabaseobject,theOracle
Databaseworksinthebackgroundtorecordthatobject’snameandstructureand
makesthatinformationavailabletoyoubywayofthedatadictionary.Every
objectcreatedbyeveryuserisdocumentedwithinthedatadictionary.
Structure
Thedatadictionaryconsistsoftablesandviewsthatareownedbytheuser
accountSYS.Asowner,SYShasfullprivilegesoverthesetablesandviews.No
usershouldeveralterdataownedbySYSorelsetheintegrityofthedatabase
maybecompromised.
TheSYSaccountisoneofafewpowerfulaccountsthatcomeswithevery
implementationoftheOracledatabase.TheSYSaccountissomethingofa
“superuser”account,withmasterprivilegestovirtuallyeverythinginthe
database.Generally,nodeveloperusestheSYSaccountforanythingother
thandatabasesystemmaintenance,andoftenit’sthedatabase
administrator(DBA)whopossessesexclusiveaccesstotheSYSaccount.
Alldatadictionaryinformationisstoredintables,butmuchofthatdatais
presentedtousersthroughviews.Inotherwords,usersgenerallydon’tgetdirect
accesstothetablesofthedatadictionary;theygetaccesstotheviews,which
providesomewhatlimitedaccessinordertoprotecttheintegrityofthedata
dictionary.
Inaddition,manydatadictionaryobjectsarerenamedviapublicsynonyms,
andthosearethenamesbywhichyouknowthedata.Inotherwords,thereare
multiplelevelsofabstractionthatseparateusersfromtheunderlyingdata.No
matter—theabilitytoreadtheinformationinthedatadictionaryisagreatasset
toallSQLprofessionals.
EveryDDLstatementthatisissuedthroughoutthedatabasecausesan
automaticupdatetothedatadictionary.ThatupdateishandledbytheOracle
systemandappliedtothebasetablesthatformthefoundationofthedata
dictionary.Usersdonotexplicitlyupdateanyinformationinthedictionary.
(Note:Thereisoneexceptiontothis.Usersmayoptionallychoosetoadd
comments,whichyou’llexplorelaterinthischapter.)
Asofthiswriting,therearemorethan2,000viewsinthedatadictionary.One
inparticularisagoodstartingpoint:DICTIONARY.Thisviewcontains
informationabouttheviewsthatcomposethedatadictionary.Itincludesthe
nameofeachview,alongwithabriefexplanationofeachview.You’lllookatit
abitlater.
TheUSER_TABLESviewcontainsinformationaboutthetablesownedby
thecurrentuseraccount.Inotherwords,nomatterwhichaccountyouloginto,
youcanquerytheUSER_TABLESviewandgetdetailedinformationaboutthe
tablesownedbywhateveraccountyouareloggedinwith.
AfulldescriptionoftheUSER_TABLESviewwouldshowthatitconsistsof
over50columns.Someofthecolumnsinclude
TABLE_NAMEThenameofthetable.
STATUSIndicateswhetherthetableiscurrentlyvalidandtherefore
availableforuse.
ROW_MOVEMENTIndicateswhetherROWMOVEMENThas
beenenabledforthetable.(SeeourdiscussioninChapter10aboutthe
FLASHBACKTABLEstatementformoreinformationaboutenabling
anddisablingROW_MOVEMENT.)
AVG_ROW_LENTheaveragelengthoftherowscurrentlystored
inthetable.
Thesearejustsomeoftheseveraldozencolumnsthatareinthe
USER_TABLESview.
Whatifyouwanttoseeinformationabouttablesotherthanyourown?Well,
itjustsohappenstherearetwootherviewsinthedatadictionarythathave
almosttheidenticalsetofcolumnsasUSER_TABLES.
ALL_TABLESShowsallthesametableinformationbutfortables
towhichthecurrentuserhasprivileges,regardlessofowner
DBA_TABLESShowsallthesametableinformationbutforallthe
tablesintheentiredatabase,regardlessofownerortableprivileges
Theseothertwoviewsalsohaveanadditionalcolumn:
OWNERTheownerofthetableinquestion
Andthatmakessense—thereisnoneedforOWNERintheUSER_TABLES
viewsincethatviewshowsinformationaboutonlyoneowner,namely,the
currentowner.
Thisnamingpatternofusingoneofthesethreeprefixes(USER_,ALL_,
DBA_)isapatternthatisusedthroughoutthedatadictionary.Manyofthedata
dictionaryviewsthatstoreinformationaboutobjectsinthedatabasehavenames
thatstartwithoneofthesethreeprefixes.Anoverviewofthesedatadictionary
viewsispresentedinTable12-1.Asyoucanseefromthetable,thevastmajority
ofdatadictionaryviewshaveaprefixofUSER_,ALL_,orDBA_.Asetofthree
viewsthathavetheUSER_,ALL_,andDBA_prefixandsharethesamesuffix,
suchasTABLES,drawtheirdatafromasingledatadictionarytable.For
example,USER_CONSTRAINTS,ALL_CONSTRAINTS,and
DBA_CONSTRAINTSsharethesamedatadictionarytable.Table12-2shows
themostcommonUSER_viewsyou'lluseonaregularbasis.
TABLE12-1 PrefixesofSomeoftheDataDictionaryViews
TABLE12-2 SelectedDataDictionaryViewsShowingObjectsOwnedby
theCurrentUser
NotethatpublicsynonymsarenotlistedintheUSER_SYNONYMSview,
whichshowsonlyprivatesynonyms.Evenifyou’reloggedintoauseraccount
thatcreatedaparticularpublicsynonymobject,you’llstillnotfinditlistedin
theUSER_SYNONYMSview.Instead,you’llfinditinALL_SYNONYMSand
DBA_SYNONYMS.
DynamicPerformanceViews
Table12-1includesreferencestoasetofviewsthatbeginwiththeprefixesV_$
andGV_$.Thesearedefinedasthedynamicperformanceviewsandtheglobal
dynamicperformanceviews.
Dynamicperformanceviewsdisplayinformationaboutcurrentdatabase
activityinrealtime.Theyreceivedatadynamicallyfromthedatabasethrough
mechanismsthatgobeyondthescopeofthisbook.Forourpurposes,it’s
importanttoknowthattheyaremaintainedautomaticallybythesystemandare
availableforquerying—withsomelimitations.
ThedynamicperformanceviewsstartwiththeprefixV_$.Therearepublic
synonymscreatedforeachoftheviews,andtheyhavesimilarnamesbutbegin
withtheprefixV$.
Simplequeriesondynamicperformanceviewsareaccepted,butcomplex
queries,withorwithoutjoins,requiresomespecialattention.Oracleformally
recommendsthatthedynamicnatureoftheseviewsdoesnotguaranteeread
consistencyforanythingotherthanthesimplestofsingle-viewqueries,soit’s
advisedthatyouperformcomplexjoinsand/orqueriesby
Creatingasetoftemporarytablestomirrortheviews
Copyingthedataoutoftheviewsandintoasetoftemporarytables
Performingthejoinonthetemporarytables
Thisway,you’llavoidgettingbadresultscausedbyalackofreadconsistency.
Someofthedynamicperformancesynonyms(whichpointtoviewsthatpoint
totables)includethefollowing:
V$DATABASEIncludesinformationaboutthedatabaseitself,
includingthedatabasename,thedatecreated,thecurrentoperating
systemplatform,andmuchmore
V$INSTANCEIncludestheinstancename,thehostname,the
startuptime,andmuchmore
V$PARAMETERThecurrentsettingsforsystemparameters,such
asNLS_LANGUAGE,NLS_DATE_LANGUAGE,NLS_CURRENCY,
NLS_TIME_FORMAT,NLS_TIME_TZ_FORMAT,
NLS_TIMESTAMP_TZ_FORMAT,SQL_TRACE,andmuchmore
Remember,onlysimplequeriesarerecommendedwhenqueryingthe
V$(v-dollar)viewsdirectly.
V$SESSIONManycurrentsettingsforeachindividualusersession,
showingactiveconnections,logintimes,machinenamesthatusersare
loggedinto,thecurrentstateoftransactions,andmuchmore
V$RESERVED_WORDSCurrentlistofreservedwords,including
informationindicatingwhetherthekeywordisalwaysreserved,andif
not,underwhatcircumstancesitisreserved
V$OBJECT_USAGEUsefulformonitoringtheusageofINDEX
objects
V$TIMEZONE_NAMESIncludestwocolumns:TZNAME,which
istimezoneregion,andTZABBREV,whichisthetimezone
abbreviation
ReadingComments
Thedatadictionaryisrichwithcommentsthathelpdescribetheintentofthe
variousviewsofthedatadictionaryandthecolumnswithinthem.Inadditionto
thecommentsthatareprovidedintheDICTIONARYviewforeachofthe
individualdatadictionaryviews,youcanviewcommentsaboutthecolumns
withinthoseviewsoraboutanyobjectstoredanywhereinthedatabase:
ALL_TAB_COMMENTSDisplayscommentsforallobjectsinthe
database
ALL_COL_COMMENTSDisplayscommentsforallcolumnsof
alltablesandviewsinthedatabase
Sayyou’relookingatadatadictionaryviewlikeUSER_SYNONYMSand
youwanttolearnmoreaboutitscolumns.Here’saquerythatwillhelpyou:
That’sthequery;herearetheresults:
Asyoucansee,we’reusingthedatadictionarytostudythedatadictionary.
Theright-sidelistingunderCOMMENTSishelpfulindescribingthecontentsof
theviewinthedatadictionary.Youcanusethistechniquetoinspectallofthe
contentsofthedatadictionary.
AddingComments
Youcanaddyourowncommentstothedatadictionarytoaddnotesand
descriptionsaboutthetablesandcolumnsyoucreate.TheCOMMENT
statementiswhatweusetoaddcommentstothedatadictionaryforaparticular
databaseobject.Itssyntaxisasfollows:
where:
objectTypeisoneofthekeywordsTABLE,COLUMN,orsome
otherobjectsthatarenotsubjectsofthecertificationexam,suchas
INDEXTYPE,OPERATOR,MATERIALIZEDVIEW,andothers.
fullObjectNameisthenameoftheobjectforwhichyouwanttoadd
acomment.Ifit’saTABLE,namethetable.Butifit’sacolumn,usethe
TABLE.COLUMNsyntax.
c1isthefulltextofthecommentyouwanttoadd.
Whenyouaddacommenttothetable,thecommentwillbedisplayedinthe
datadictionaryviewsUSER_TAB_COMMENTS,ALL_TAB_COMMENTS,
andDBA_TAB_COMMENTS.
Whenyouaddacommenttoacolumninatable,thecommentwillbe
displayedinthedatadictionaryviewsUSER_COL_COMMENTS,
ALL_COL_COMMENTS,andDBA_COL_COMMENTS.
Forexample,let’ssaywewanttoaddacommenttothedatadictionaryabout
thePORTStable.Here’sanexample:
Toseetheresults,youcouldusethisquery:
Here’sanexampleofaddingacommenttoatable’scolumn:
Youcan’treallydropacommentfromthedatadictionary.Instead,you
changeittoablank,likethis:
Let’stakealookatsomeusefulexamplesofthedatadictionaryinaction.
DICTIONARY
TheDICTIONARYviewisagreatstartingpointforanyinvestigationofthe
datadictionary.IfweDESCRIBEtheview,wegetthefollowing:
TherearejusttwocolumnsintheDICTIONARYview,butnotethatthe
secondcolumncanpotentiallyholdagreatdealofinformation.
InourcurrentinstallationoftheOracledatabase(usingacontainerdatabase),
we’reshowing3,228entriesinthisview.Youcanrunasimplequerytolistallof
itscontentsinyourownuseraccount.
Theoutputistoomuchtolisthere,andit’sgenerallythesamehereasitwill
beonyourownOracledatabaseimplementation,dependingonwhichversion
you’reusing.Youmightwanttorunthisqueryinyourownuseraccountand
reviewitsoutput.Ifyou’relookingforsomethingspecific,suchasanythingthat
addressesindexobjects,youmighttryaquerylikethis:
ThatquerywilllocateanythingintheDICTIONARYtablethatmentions
“index”inthecomments.Theresultwillincludethenameofthedatadictionary
viewthatlistsalloftheindexes,theonethatlistsallofthecolumnsuponwhich
anindexisbased,andsoon.
Then,ifyoulocateaparticularentryinthedictionaryyouwanttoknowmore
about—forexample,USER_DEPENDENCIES—youcanrunthefollowing
querytogetcommentsonthatparticularviewanditscolumns:
Thosequeriesshouldhelptozeroinonhelpfulinformationinthedata
dictionary.
IdentifyingaUser’sOwnedObjects
Thereareavarietyofdatadictionaryviewsfromwhichyoumightgatherdata
aboutyourownuseraccount’sobjects.Twoviewsinparticularareagood
startingpoint:USER_CATALOGandUSER_OBJECTS.
USER_CATALOG
TheUSER_CATALOGviewdisplaysasummarylistingoftables,views,
synonyms,andsequencesownedbytheuser.SeeFigure12-1foradescription
oftheview.
FIGURE12-1 DescriptionoftheUSER_CATALOGdatadictionaryview
Here’sasamplequerytogetaquickoverviewofwhataparticularuser
accountmayown:
ThereareonlytwocolumnsinUSER_CATALOG;theyareTABLE_TYPE
andTABLE_NAME,whereTABLE_NAMEisactuallythenameofthetable,
view,sequence,orsynonymobject.
AsynonymforUSER_CATALOGisCAT.
Besureyouhaveatleastabasicworkingknowledgeofeachofthe
datadictionaryviewsthattrackthebasicobjectsinthedatabase—tables,
views,sequences,synonyms,sequences,constraints—andthedifference
foreachwithregardtotheUSER_,DBA_,andALL_prefixes.
USER_OBJECTS
TheUSER_OBJECTSviewcontainsinformationaboutallobjectsownedbythe
user.AsynonymforUSER_OBJECTSisOBJ.SeeFigure12-2fora
description.
FIGURE12-2 DescriptionoftheUSER_OBJECTSdatadictionaryview
InspectingTablesandColumns
TheUSER_TABLEStable(synonymTABS)ishelpfulforinspectingtable
metadata,asisitscompanionUSER_TAB_COLUMNS(synonymCOLS).This
sectionwilllookatUSER_TAB_COLUMNSinparticular.
Let’sgetcolumninformationforthetableshowninFigure12-3.
FIGURE12-3 DiagramoftheINVOICEStable
HereisaSELECTstatementthatwillpullinformationfromthedata
dictionaryaboutthecolumnsofINVOICES:
Here’stheoutput:
Note:Fortherecord,theprecedingSELECTstatementisn’ttotallyperfect.It
onlyaddressesdatatypesofDATE,NUMBER,andVARCHAR2,andinthe
eventaNUMBERdatatypehasaprecisionbutnoscale,theformattingwon’t
comeoutlookingquiteright.Thepointhereistoillustratethesortofqueryyou
mightwanttodoinordertoextractdatadictionaryinformationfromthe
database.
CompilingViews
Oneofthemanyusefultasksyoucanaccomplishwiththedatadictionaryisto
checkforthestatusofaviewthatyou’vecreated.RememberfromChapter10
thataviewisanamedquerybasedonatableandthataftertheviewhasbeen
created,ifthetableisalteredforanyreason,youmayhavetorecompilethe
view.Forexample,ifatable’sstructureisaltered,suchasbyachangetoa
column’sdatatypeorperhapsifacolumnisdroppedfromthetablealtogether(a
columnthatisusedbytheview),thenitmaychangethestatusoftheviewto
INVALID.
Youcancheckthedatadictionary’sUSER_OBJECTSviewtodeterminethe
statusofanyofyourviews,likethis:
Inourcase,hereistheoutput:
Sonowweknowweneedtorecompiletheseviews.SeeChapter10fordetails
abouthowtorecompileaview.
Thedatadictionarycontainsalotofinformationaboutviews,includingthe
queryuponwhichtheviewisbased,whichcanbefoundintheUSER_VIEWS
viewanditsTEXTcolumn.Here’saqueryonthedatadictionarythatasksfor
thequerythatwasusedtocreatetheviewVW_EMPLOYEES:
Hereistheoutput:
Insummary,allthemetadataforyourdatabaseisstoredinthedatadictionary
andisavailablefordisplayandinspection.
CheckingPrivileges
PrivilegesarediscussedinChapter14,whenwediscussuseraccess.Fornow,
notethatprivilegescanbeinspectedusingthefollowingviews:
USER_SYS_PRIVSSystemprivilegesgrantedtothecurrentuser
USER_TAB_PRIVSGrantedprivilegesonobjectsforwhichthe
useristheowner,grantor,orgrantee
USER_ROLE_PRIVSRolesgrantedtothecurrentuser
DBA_SYS_PRIVSSystemprivilegesgrantedtousersandroles
DBA_TAB_PRIVSAllgrantsonobjectsinthedatabase
DBA_ROLE_PRIVSRolesgrantedtousersandroles
ROLE_SYS_PRIVSSystemprivilegesgrantedtoroles
ROLE_TAB_PRIVSTableprivilegesgrantedtoroles
SESSION_PRIVSSessionprivilegesthattheusercurrentlyhasset
Eachcanbeinspectedbytheusertodeterminethecurrentstateofprivileges
androles.SeeChapter14forafulldiscussionofuseraccess,privileges,and
roles,includingsamplequeriesofthesedatadictionaryviews.
InspectingConstraints
TheUSER_CONSTRAINTSviewisoneofthemoreusefulviews.Here’sa
queryyoumightruntocheckthecurrentstateofconstraintsonatable
CRUISES:
Hereisanexampleoftheoutput:
TheoutputlistsalltheconstraintsontheCRUISEStable.We’reseeingfour:
oneprimarykeyandthreeforeignkeys.
HowdoyouknowwhichconstraintisaPRIMARYKEYandwhichisa
FOREIGNKEY?TheanswerisbytheCONSTRAINT_TYPEcolumn.Someof
thepossibleentriesintheCONSTRAINT_TYPEcolumnare
P=PRIMARYKEY
R=FOREIGNKEY(theRisfor“referentialintegrity”)
U=UNIQUE
C=CHECKorNOTNULLconstraint
Note:Thereareadditionalconstraints;thesearebeyondthescopeofthe
exam.
TheDELETE_RULEcolumnshowswhetheraforeignkeyconstraintwas
createdwithONDELETECASCADEorONDELETESETNULL.
TheSEARCH_CONDITIONcolumnisparticularlyusefulforinspecting
CHECKconstraintcriteria.Here’sanexample:
Takenoteoftheconstraintswithunexpectedvaluesfor
CONSTRAINT_TYPE:RforFOREIGNKEY,andCforNOTNULLas
wellasforCHECK.
Thedatadictionaryprovidesadditionalinformationaboutconstraintsinthe
USER_CONS_COLUMNSdatadictionaryview.Thatviewcontainsallthe
informationaboutwhichcolumnsinCRUISESareconstrainedandwhatthe
namesareofthereferencedtablesandcolumnsthatmakeuptheFOREIGN
KEYconstraints.
FindingColumns
HereisonetypeofdatadictionaryqueryIfinduseful:
That’saquerythatlooksforalltablesinthecurrentuseraccountthathappen
tohaveaparticularcolumn—inthiscase,onenamedEMPLOYEE_ID.Ifindit
helpfultosearchforparticularcolumnnamesacrossalltablesorviewsina
givenuseraccount(i.e.,aschema).
Therearemanyhelpfulsoftwaretoolsavailablethatwillextractdata
dictionaryinformation—suchascomments—andprovideanicepoint-and-
clickinterfacetomakeiteasytonavigate.That’sallveryhelpful.But
sometimesyou’llfindyourselfinasituationwhereyousimplydon’thave
accesstothosetools.Youmightevenrealizethataparticularapplication
you’redevelopingcouldbenefitbyprogrammaticallyaccessingdatadictionary
informationbywayofSQLstatementstodrawdataintoyourapplicationfor
someprojectrequirement.Thepointisthatthedatadictionaryisa
certificationexamobjectiveforagoodreason;acomfortableunderstandingof
itsinformationandanabilitytonavigateiteasilyisimportantforanyserious
SQLprofessional.
CERTIFICATIONSUMMARY
Thedatadictionaryisapowerfultool.Itconsistsofaseriesoftablesandviews
thatareautomaticallymaintainedbytheOraclesystemtodocumentthestateof
everyobjectinthedatabase.WheneveraDDLstatementisexecuted,thedata
dictionaryisupdatedinsomefashion.
Thedatadictionaryisoftenreferredtoasmetadata,atermthatmeans“data
aboutdata.”Thedatadictionarycontainsinformationaboutthedatabaseobjects
youcreate—theirstructures,names,status,andmore.
TheSYSaccountownsthedatadictionary’sbasetables,whichcannotbe
changeddirectlybyusers.Instead,allofthetableshaveviewsand,insome
cases,publicsynonymsthathavebeencreated,anditisthesetheuseraccesses
inaread-onlymode.
Manyofthedatadictionaryviewsfollowaprefixpatternthatindicatesthe
contentsoftheview.ViewswithaprefixofUSER_showdataaboutobjects
ownedbytheuseraccessingtheview.ALL_istheprefixforobjectsthatexist
anywhereinthedatabasetowhichthecurrentuserhasaccess.DBA_isthe
prefixforviewsthatshowdataaboutallobjectsinthedatabase,regardlessof
whoownsthemorwhatprivilegesmaybegrantedtothem.
Informationinthedatadictionaryincludesthenamesoftablesandtheir
columns,includingeachcolumn’sdatatype,alongwithitsprecision,scale,
and/orlengthwhereapplicable.Allofthedatabaseobjectsarelistedinthedata
dictionary,includingalltables,views,indexes,sequences,constraints,andmore.
ViewsthathaveaprefixofV$orsomevariationaredynamicperformance
viewsandshowreal-timedatabaseperformanceinformation.Oraclecannot
guaranteethereadconsistencyoftheseviews,soit’srecommendedthatfor
dynamicperformanceviewsyoulimityouraccesstosingle-tablequeries.If
morecomplexqueriesand/orjoinsarerequired,youareadvisedtofirstcopy
dataoutoftheviewsintoyourowntemporarytablesandthenquerythosetables
forbetterresultsanddataintegrity.
Youcanaddcommentstotheentriesinthedatadictionaryforyourown
tablesandcolumnsusingtheCOMMENTstatement.Youcannotdelete
commentsbutinsteadupdatecommentswithablankstring.
Thedatadictionarycanbeusedtoperformavarietyofusefultasks,suchas
obtaininginformationabouttimezones,determiningwhetheraviewrequires
recompilation,identifyinganyprivilegesthatarecurrentlygranted,andmuch
more.
✓TWO-MINUTEDRILL
QueryVariousDataDictionaryViews
Thedatadictionaryismadeoftablesthatstoredataaboutthe
database.
Thedatadictionarycontainsthemetadataforyourdatabase.
Itcontainsinformationabouttables,views,constraints,indexes,
sequences,roles,privileges,andanyandallotherobjectsyoumight
createinthedatabase.
Itkeepstrackofalltheusersinthedatabaseandwhichuser
accountownswhichobjects,whohasprivilegesonwhichobject,the
statusofeachobject,andmore.
Oracleautomaticallyupdatesandmaintainsthedatadictionary
viewswitheachDDLstatementexecutedthroughoutthedatabase.
ThedatadictionaryviewsthatbeginwiththeprefixUSER_contain
informationaboutobjectsownedbytheuseraccessingtheview.
TheALL_prefixindicatesadatadictionaryviewthatcontains
informationaboutobjectsthatmightbeownedbyanyuserinthe
databasebuttowhichtheaccessinguserhasprivileges.
TheDBA_prefixisaffixedtoallviewsthatcontaindataaboutall
objectsinthedatabase.
TheV$orGV$prefixidentifiesviewsthatarepartofthesetof
dynamicperformancetablesandviews,whichshowreal-time
performancedataaboutthedatabase.
Most(butnotall)ofthedatadictionaryviewsarestoredwith
commentsthatprovidebriefdescriptionsabouteachviewandwhatit
contains;manyofthecolumnsoftheviewsalsohavecomments.
Youcanaddcommentsofyourownalongsidethedatadictionary
recordforyourownobjectsthatyou’vecreated.
TheCOMMENTstatementishowyoustoreacommentinthedata
dictionaryforanytableyouownandforitsassociatedcolumns.
TheDICTIONARYviewisagreatstartingpointforfindingwhat
youmightbelookingforinthedatadictionary.
TheUSER_CATALOGviewcontainsasummaryofinformation
aboutsomeofthemajorobjectsownedbyyouruseraccount.
TheUSER_OBJECTSviewissimilartoUSER_CATALOGbut
withmuchmoreinformation.
Youcangetafulllistingfromthedatadictionaryforyourtables;
theircolumns;andassociateddatatypes,lengths,precision,andscale.
Thestatusofobjectsisalsostored;forexample,thedatadictionary
flagsviewsthatareinvalidandmightneedrecompilation.
Allrolesandprivilegesofallusersonallobjectsarestored
somewhereinthedatadictionary.
Ifyouhavethenameofacolumnandaren’tsurewhichtableit
mightbepartof,usetheALL_TAB_COLUMNSdatadictionaryview.
SELFTEST
Thefollowingquestionswillhelpyoumeasureyourunderstandingofthe
materialpresentedinthischapter.Chooseonecorrectanswerforeachquestion
unlessotherwisedirected.
QueryVariousDataDictionaryViews
1.Oneplacetogetamasterlistofalltheviewsthatformthedata
dictionaryis:
A.DICTIONARY
B.DATA_DICTIONARY
C.CATALOG
D.USER_CATALOG
2.Youaretaskedwithqueryingthedatadictionaryviewthatlistsonly
thosesequencestowhichyoucurrentlyhaveprivilegesbutdon’t
necessarilyown.Todothis,youlogintoyourownuseraccountandquery
thedatadictionaryviewcalled:
A.ALL_SEQUENCES
B.DBA_SEQUENCES
C.USER_SEQUENCES
D.USER_PRIV_SEQUENCES
3.Whichofthefollowingactionswillnotcausethecontentsofthedata
dictionarytobechangedinsomeway?
A.Createanewtable
B.Modifythedatatypeofanexistingcolumn
C.ExecuteavalidCOMMENTstatement
D.Noneoftheabove
4.Thedatadictionaryisownedby:
A.PUBLIC
B.SYS
C.SYSTEM
D.Eachindividualuser
5.Youcanaddyourowncommentstothedatadictionarywiththe
COMMENTstatementusingwhichofthefollowing?(Choosetwo.)
A.INDEX
B.COLUMN
C.SEQUENCE
D.TABLE
6.Youneedtogetinformationaboutcolumnsinatableyoudonotown,
nordoyouhaveprivilegestoit.Whichviewcanyouquerytogetthis
information?
A.DBA_TAB_COLUMNS
B.ALL_TAB_COLUMNS
C.ALL_COLUMNS
D.Can’tbedone
7.Whichamongthefollowingisconsideredanacceptablequerywith
V$DATAFILE?
A.Ajoinwithtwootherobjectsinthedatadictionary
B.AcomplexGROUPBYwithmultiplelevelsofaggregation
C.Aquerythatdisplaysrowsfromthetablewithnojoins
D.Alloftheabove
8.Youaretaskedwiththejobofaddingacommenttothedatadictionary
toaccompanythecolumnPIERinthetableMARINA.Whichofthe
followingwillexecutesuccessfully?
A.COMMENTONCOLUMN(MARINA.PIER)IS'Numberof
piers';
B.COMMENTONCOLUMNMARINA.PIERIS'Numberof
piers';
C.COMMENTONCOLUMNMARINA(PIER)IS'Numberof
piers';
D.COMMENTONTABLECOLUMNMARINA.PIERIS
'Numberofpiers';
9.NowyouhavechangedthepurposeofthePIERcolumninthe
MARINAtableandwanttoremovethecommentyoujustcreatedinthe
previousquestion.Whichofthefollowingstatementswillremovethe
comment?
A.COMMENTONCOLUMNMARINA.PIERDROP;
B.COMMENTONCOLUMNMARINA.PIERISNULL;
C.COMMENTONCOLUMNMARINA.PIERSETUNUSED;
D.COMMENTONCOLUMNMARINA.PIERIS‘’;
10.Whenyou’relookingforaparticularbitofdataandyou’renotsurewherein
thedatadictionaryitmightbe,agoodstartingpointis:(Choosethebest
answer.)
A.SELECT*FROMV$DATABASE;
B.SELECT*FROMGV_$START_HERE;
C.SELECT*FROMDICTIONARY;
D.SELECT*FROMV$RESERVED_WORDS;
11.TheUSER_CONSTRAINTSviewinthedatadictionarylistsFOREIGN
KEYconstraintsintheCONSTRAINT_TYPEcolumnwithwhichofthe
followingsingle-letterabbreviations?
A.K
B.R
C.F
D.G
12.Youaretaskedtoworkwithaview.Theview’sunderlyingtablehasbeen
altered.Whatinformationcanthedatadictionaryprovideatthispoint?
(Chooseallcorrectanswers.)
A.Thestatusoftheviewsothatyoucandeterminewhetherthe
viewrequiresrecompilation
B.Thecurrentstateofthetable
C.Thequerythatwasusedtocreatetheview
D.Thenamesofcolumnsintheunderlyingtable
13.Thetermmetadatameans:
A.Dataaboutdata
B.Globaldatathatisaccessiblethroughoutthedatabase
C.Datathatisautomaticallyupdatedandmaintainedbythe
databasesystem
D.Distributeddata
14.WhichofthefollowingdatadictionaryviewsdoesnothaveanOWNER
column?
A.USER_TABLES
B.ALL_INDEXES
C.DBA_CONS_COLUMNS
D.Alloftheabove
15.IfanALTERTABLE...DROPCOLUMNstatementisexecutedagainstan
underlyingtableuponwhichaviewisbased,thestatusofthatviewinthe
datadictionarychangesto:
A.COMPILE
B.INVALID
C.ALTERED
D.FLAG
SELFTESTANSWERS
QueryVariousDataDictionaryViews
1. A.DICTIONARY.YoucanruntheDESCDICTIONARYcommand
toseethetwocolumnsthatformDICTIONARYandthenqueryitfor
additionalinformation.
B,C,andDareincorrect.Thereisnosystemviewcalled
DATA_DICTIONARY.CATALOGisalsoincorrect;thereisaCATALOG
viewincludedwithOracleforbackwardcompatibilitywithversion5,but
Oracleofficiallydiscouragesitsuse.Ontheotherhand,thereisaviewin
thedatadictionarycalledUSER_CATALOG;it’sarecommendedresource
forfindinginformationabouttables,views,andotherdatabaseobjectsthe
currentuserowns.
2. A.ALL_SEQUENCESwilllistanysequencesinthedatabase,
regardlessofowner,towhichyouraccounthasbeengrantedaccess.
B,C,andDareincorrect.DBA_SEQUENCESwilllistallsequencesin
thedatabase,regardlessofwhoownsthemandregardlessofwhohas
privilegesonthem.USER_SEQUENCESwilllistonlythosesequencesthat
youruseraccountcurrentlyowns.Thereisnoviewcalled
USER_PRIV_SEQUENCES.
3. D.Noneoftheabove.
A,B,andCareincorrect.Allofthesewillenactsomesortofchangeto
theinformationinthedatadictionary.AnyDDLthatcreatesormodifies
objectswillupdatetheobjectlistingsinthedatadictionary.The
COMMENTstatementaddstothedatadictionaryacommentabouta
particularobject.
4. B.SYSistheownerofthedatadictionary.
A,C,andDareincorrect.NeitherPUBLICnorSYSTEMownsthedata
dictionary,althoughbotharevalidusersinthesystem.
5. BandD.TABLEandCOLUMNobjectsaresupportedbythe
COMMENTstatement.
AandCareincorrect.Youcannotaddcommentsofyourowntothedata
dictionaryforanINDEXorSEQUENCEobject.
6. A.DBA_TAB_COLUMNSistheviewthatcontainsinformation
aboutcolumnsintables,andbecauseoftheDBA_prefix,youknowit
containsinformationabouttablesandcolumnsthatexistanywhereinthe
database,regardlessofownerorprivilegesgranted.
A,C,andDareincorrect.ALL_TAB_COLUMNSwouldbecorrectif
youwerelimitingyoursearchtotablesandcolumnstowhichyou’vebeen
grantedprivilegesbecausethatiswhattheALL_prefixindicates.Thereis
noALL_COLUMNSviewinthedatadictionary.
7. C.TheV$prefixindicatesthatV$DATAFILEisapublicsynonym
foradynamicperformanceview,forwhichOracleCorporationdoesnot
guaranteereadconsistency.Therefore,youarerecommendedtolimityour
directaccessofV$objectstosimplequeries.
A,B,andDareincorrect.OracleCorporationofficiallyadvisesagainst
usinganyoftheV$objectsincomplexqueriesand/orjoins.
8. B.ThecorrectsyntaxistousethekeywordsCOMMENTON
COLUMN,followedbythetablenameandcolumnname,separatedbya
period,andthekeywordIS,followedbythestring.
A,C,andDareincorrect.ParenthesesarenotpartoftheCOMMENT
statement.ThekeywordTABLEisusedonlywhenaddingacommenttoa
table.
9. D.Therereallyisn’tastatementtoexplicitlydropacommentor
deleteit.Thepracticeistooverwritetheoldcommentwithanemptystring.
A,B,andCareincorrect.Noneoftheseoptionsisavalidstatement.
Theycontainbitsandpiecesofvalidreservedwordsfromotherstatements
butdonotapplytoCOMMENT.
10. C.TheDICTIONARYviewsummarizesthenamesoftablesandviewsin
thedatadictionary,alongwithdetailedcommentsabouteachone.
A,B,andDareincorrect.TheV$DATABASEand
V$RESERVED_WORDSobjectsarevalidpublicsynonymsfordata
dictionaryviews,butthesearepartofthesetofdynamicperformance
tablesandnotgoodforgettinganoverviewofthedictionaryasastarting
point.ThereisnosuchobjectintheOracledatadictionarycalled
GV_$START_HERE.
11. B.Ristheanswer.Rstandsfor“referentialintegrity”andindicatesthe
presenceofaFOREIGNKEYconstraintintheCONSTRAINT_TYPE
columnoftheUSER_CONSTRAINTSdatadictionaryview.
A,C,andDareincorrect.It’snotKorG.Andyou’dthinkitwouldbeF,
butit’snot.Rmakessense,though,whenyouthinkaboutit.
12. A,B,C,andD.Thedatadictionarycanassistwithalloftheanswers
listed.
Noneareincorrect.
13. A.Metadatais“dataaboutdata.”
B,CandDareincorrect.
14. A.USER_TABLESdoesnothaveorneedanOWNERcolumnsincethe
viewonlypresentsasetoftablesownedbytheuseraccessingthem.
B,C,andDareincorrect.Evenifyouhaven’tlookedindetailatthese
views,youcanrelyonthefactthatviewsthatstartwithALL_andDBA_
haveacolumnshowingOWNERinformationsincetheycontain,by
definition,objectsownedby—potentially—morethanoneuser.
15. B.ItchangestoINVALID.Recompilingtheviewcouldrestorethestatus
oftheviewtoVALID.
A,C,andDareincorrect.
13
ManipulatingLargeDataSets
T
CERTIFICATIONOBJECTIVES
13.01DescribetheFeaturesofMultitableINSERTs
13.02MergeRowsintoaTable
✓Two-MinuteDrill
Q&ASelfTest
hischapterlooksatfeaturesandoperationsthatareusefulforworkingwith
largegroupsofdata.Wecovertwooperationsinparticular.Thefirstconsists
ofadditionalfeaturestotheINSERTstatementthatenableyoutouseone
INSERTstatementtoaddmultiplerowsofdatatoagiventableortoseveral
tables,withandwithoutconditionallogic.Laterinthechapterwe’llcoverthe
SQLstatementMERGE,whichisapowerfulstatementthatenablesyouto
combinethefeaturesofmultipleDataManipulationLanguage(DML)
statementsintoasingleSQLstatement.
CERTIFICATIONOBJECTIVE13.01
DescribetheFeaturesofMultitableINSERTs
ThemultitableINSERTstatementisavariationontheINSERTstatementsyntax
you’vealreadyseen.AmultitableINSERTstatementrepeatstheINTOclauseof
theINSERTstatementtoinsertdataintomorethanonetable.EachINTOclause
appliestojustonetable,butbyrepeatingtheINTOclause,youcanadddatato
multipletables.ThemultitableINSERTmusthaveasubquerytoselectrowsfor
inserting.
MultitableINSERTstatementscanaccomplishavarietyoftasks,including
thefollowing:
Querydatafromonetableandinsertthedataintomultipletables
withconditionallogic,suchastransformingdataintoaseriesofarchive
tables.
Exchangedatabetweentwosimilarsystemsofdifferent
requirements—perhapsbetweenatransaction-basedapplicationanda
datawarehouseoptimizedforanalysis.
Supportlogicalarchivingatanylevelofdetailwithlogicaldecision
pointsembeddedintheINSERTstatements.
IntegratecomplexquerieswithGROUPBY,HAVING,setoperators,
andmore,allwhilemovinganynumberofrowsdynamically,
distributingoutputintomultipledatatargets,andprogramminglogical
decisionpointstocontroldatadistribution.
Transformdatathatisstoredinrowsandlevelsintoacross-
tabulationoutput,thetypeyouwouldtypicallyseeinaspreadsheet
application.
TherearetwogeneraltypesofmultitableINSERTstatements:unconditional
andconditional.
UnconditionalmultitableINSERTstatementsprocesseachofthe
INSERTstatement’soneormoreINTOclauses,withoutcondition,forall
rowsreturnedbythesubquery.
ConditionalmultitableINSERTstatementsuseWHENconditions
beforeINTOclausestodeterminewhetherthegivenINTOclause(or
clauses)willexecuteforagivenrowreturnedbythesubquery.Inother
words,foreachrowreturnedbytheINSERTstatement’ssubquery,each
WHENcondition’sexpressionisconsideredandevaluatestoeitheratrue
orfalsecondition.Iftrue,theassociatedINTOclause(orclauses)will
execute.Iffalse,itwillnot.Finally,anoptionalELSEclausecaninclude
analternativeINTOclausethatcanbeexecutedifnoneoftheWHEN
conditionsistrue.
Let’slookattheoverallsyntaxforthemultitableINSERTstatement.First,
we’llexamineanunconditionalmultitableINSERTstatement.Thesyntax
repeatstheINTOstatementfromonetomanytimesasrequired.
TheunconditionalmultitableINSERTstatementsyntaxjustshownassumes
thefollowing:
ThekeywordALLisrequiredinanunconditionalmultitable
INSERT.Note,however,thatwhilethepresenceofthekeywordALLis
indicativeofamultitableINSERT,itdoesn’tnecessarilyindicatethe
unconditionalmultitableINSERT,asyou’llseeinthenextsection.
FollowingthekeywordALL,theremustbeatleastoneINTO
clause.
YoucanincludemultipleINTOclauses.
EachINTOmayhaveitsownVALUESclause.
EachVALUESlistisoptional;ifomitted,theselectlistfromthe
subquerywillbeused.
ThesubqueryisaSELECTstatementthatcanstandalone.
TheconditionalmultitableINSERTstatementsyntaxissimilarbutaddsthe
WHENcondition,likethis:
Foreachrowreturnedbythesubquery,eachWHENconditionisevaluated
anddeterminedtobeeithertrueorfalse.Iftrue,thentheWHENcondition’s
associatedsetofoneormoreINTOclausesisexecuted;otherwise,processing
skipsovertheINTOclausestothenextWHENcondition.IfnoneoftheWHEN
conditionsevaluatestotrue,theELSEclauseisprocessed,anditsassociatedset
ofoneormoreINTOclausesisexecuted.
TheconditionalmultitableINSERTstatementsyntaxjustshownisusedas
follows:
Theoptionisoneoftwokeywords:ALLorFIRST.
ALListhedefaultandmaybeomitted.
FIRSTisthealternativekeyword;itindicatesthattheonlysetof
INTOclausesthatwillexecutearethosethatfollowthefirstWHEN
clausethatevaluatestotrue.
YoucanincludemultipleWHENconditions.
EachWHENconditionisfollowedbyoneormoreINTOclauses.
EachINTOmayhaveitsownVALUESclause;ifomitted,the
subquery’sselectlistmustmatchthenumberanddatatypesoftheINTO
table’scolumns.
Eachexpressionevaluatestotrueorfalseandshouldinvolveoneor
morecolumnsfromthesubquery.
Thetabandcol_listarethecomponentsoftheINSERTstatement
thatwillexecuteiftheWHENexpressionevaluatestotrue.
TheoptionalELSE...INTOclause,ifincluded,mustbelast.
ThesubqueryisrequiredandmustbeavalidSELECTstatement.
AconditionalmultitableINSERTstatementwillprocesseachrowreturnedby
thesubquery.
ThemultitableINSERTstatementalwaysusesasubquery.Asyouknow,a
subquerymayreturnanywherefromzerotomanyrows.Foreachrowreturned
byasubquery,processingdoesapassthroughthesetofWHILE…INTO
clauses.ButthewayitprocessestheWHILE…INTOclausesdiffersbasedon
whetherthekeywordALLorFIRSTisused.
IfthekeywordALLisspecified,thenalltheWHENconditionswillbe
evaluatedforeachrowreturnedbythesubquery.ForeachWHENconditionthat
evaluatestotrue,thecorrespondingINTOclause(orclauses)thatfollowthe
WHENwillbeexecuted.ForeachWHENconditionthatevaluatestofalse,the
correspondingINTOclause(orclauses)willnotbeexecuted.AlltheWHEN
conditionsareevaluatedifthekeywordALLisspecifiedatthebeginningofthe
multitableINSERTstatement.
Ontheotherhand,ifthekeywordFIRSTisused,thenforeachrowreturned
bythesubquery,WHENconditionsareevaluateduntilthefirsttrueconditionis
encountered.AssoonasaWHENconditionisdeterminedtobetrue,the
correspondingsetofoneormoreINTOclausesthatfollowstheWHENwillbe
executed.ProcessingwillthenskipovertheremainingWHENconditionsfor
thatrowofthesubquery.
Ineithersituation—INSERTFIRSTorINSERTALL—ifnoWHEN
conditionwasfoundtobetrueandiftheoptionalELSEclauseispresent,then
theELSEclause’sINTOclausewillbeexecutedfortherow.Thenprocessing
movesontothenextrowreturnedbythesubquery.
NotethatfortheconditionalmultitableINSERTstatement—whichistosay
anymultitableINSERTwithaWHENcondition—ALListhedefaultkeyword.
IfnoWHENconditionisused,thenthemultitableINSERTisunconditional,and
theALLkeywordmustbepresent.
Inotherwords,youmaynotomitthekeywordinanunconditionalmultitable
INSERT,likethis:
TheprecedingstatementshowsincorrectsyntaxsinceitomitstheALLoption
andyethasnoWHENcondition.Therefore,itissyntacticallyincorrect.
However,youmaydosomethinglikethis:
TheprecedingunconditionalmultitableINSERTcorrectlyshowstheALL
keyword.
TheconditionalmultitableINSERTallowsyoutoomitthekeyword,likethis:
MultitableINSERTstatementsrequireasubquery.
ThedefaultkeywordisALL.InallformsofthemultitableINSERT,the
subqueryisrequired;itisnotoptional.Andasisalwaysthecasewithany
INSERTthatusesasubquery,theINSERTstatementwillexecuteonceforeach
rowreturnedbythesubquery.
NotethatifanyoneINTOclausefailswithanexecutionerrorforanyone
rowreturnedbythesubquery,thentheentirestatementfailsforallrowsofthe
subquery,andnodatachangeresults.
UsetheFollowingTypesofMultitableINSERTS:
UnconditionalandConditional
Let’slookatsomeexamplesofthemultitableINSERTstatementinaction.First
upareunconditionalmultitableINSERTs,followedbyconditionalexamples.
Thenwe’llconsidertheuniqueandusefulapproachtoperformtheequivalentof
aspreadsheetpivot,acommonuseofthemultitableINSERT.
Unconditional
Aswediscussed,theunconditionalmultitableINSERTstatementomits
conditionallogic,inotherwords,theWHENclause.Forexample,considerthe
CRUISE_ORDERStable,asshowninFigure13-1.
FIGURE13-1 DiagramfortheCRUISE_ORDERStable
InadditiontotheCRUISE_ORDERStable,ourexampleusesthree
identicallystructuredtablesnamedCO_2018,CO_ELCARO,and
CO_ARCHIVED,eachwiththesamecolumnsasthetableCRUISE_ORDERS.
Thethreeidenticallystructuredtablesareusedforarchivingandanalyzingthe
CRUISE_ORDERStabledata.
HereisanexampleofavalidSQLstatementthatqueriesthe
CRUISE_ORDERStableandinsertstheoutputintoeachofourthreearchive
tables(linenumbersadded):
NotethatwehavethreeINTOclauseshere.Ifthesubqueryreturns,for
example,threerows,thentheresultofthisINSERTstatementwillbetoinsert
ninerows:threeintotheCO_2018table(line2),threeintotheCO_ELCARO
table(line6),andthreeintotheCO_ARCHIVEDtable(line10).
Asweseeintheprecedingexample,theunconditionalINSERTstatement
usesthekeywordALL(line1),followedbyoneormoreINTOclauses(lines2,
6,and10),eachofwhichspecifiesatableandthecolumnsintowhichweare
insertingdata,followedbytheVALUESlist.
TheVALUESlistcanspecifyexpressionsfoundinthesubquery’sselectlist.
Inourexample,inline4wespecifyCRUISE_ORDER_IDasthefirst
expressionintheVALUESlisttobeinsertedintotheCO_2018table.This
correspondstotheCRUISE_ORDER_IDcolumninthesubqueryselectlistin
line14.TheotherVALUESliststhatrefertoCRUISE_ORDER_ID(line8and
line12)arespecifyingthatsamecolumn.EachVALUESlistinamultitable
INSERTcanspecifyanycolumnnamesorexpressionsthatareinthesubquery
selectlist.
Ontheotherhand,thecolumnreferenceswithineachINTOlist(eachstarting
atlines2,6,and10)specifythecolumnsofthetablesnamedfortheINTO
clause.Inourexample,line2namestheCO_2018table,andtheINTOlistthat
followsonline2andline3specifiescolumnsintheCO_2018table.
You’llrecallthatinastandardINSERTstatement,thelistofvaluesinthe
VALUESexpressionlistmustmatchinnumberandindatatype(orbeabletobe
automaticallyconvertedtoamatchingdatatype)withthecolumnsspecifiedin
theINTOclause.ThesameistruehereforeachpairofINTOandVALUES
lists.
EachVALUESexpressionlistmayuseanycomplexexpressioninspecifying
thevaluetobeinsertedintoitscorrespondingtableandcolumn.Here’san
example:
Inthisexample,wearechoosingtoinsertsomevaluesthataredifferentfrom
whatthesubqueryisreturning.FortheCO_2018table,inlines2through4,we
aredefiningtheORDER_DATEvalueforallrowstobeSYSDATE,the
CRUISE_CUSTOMER_IDvaluetobetheliteralvalueof14,andtheSHIP_ID
valuetobealiteralvalueof1.FortheCO_ELCAROtable,inlines5through7,
wearegivingeachrowanORDER_DATEvaluethatis30daysbeyondthe
incomingvalueinthesubquery,andwe’reassigningthenumber15toeach
CRUISE_CUSTOMER_IDandassigning1toeachSHIP_ID.Forthe
CO_ARCHIVEDtable,inlines8through11,wearechoosingtopassthrough
valuesfromthesubqueryunchanged.
Astheexampleshows,theVALUESlistcanspecifycolumnnamesand
expressionsfromthesubquery’sselectlist,butmayalsodefineanyvalidSQL
expression.TheINTOcolumnlistmustspecifycolumnsinthetableintowhich
theINTOstatementisinsertingdata.
IftheVALUESlistisomitted,thecolumnsofthesubquerybecomethede
factoVALUESlistandthereforemustmatchthecolumnsofthecorresponding
INTOclause.By“match,”wemeanthattheymustmatchinnumberandindata
type,orbeofsuchdatatypesthatanautomaticdatatypeconversionmaybe
performed.
IfthereisnocolumnlistintheINTOclause,thesubquery’sselectlistmust
matchthecolumnsinthetableoftheINTOclause.
Conditional
ConditionalmultitableINSERTstatementsuseconditionallogictodetermine
whichINTOclauseorclausestoprocess.Eachrowthatisreturnedbythe
subqueryisprocessedthroughaseriesofoneormoreWHENconditions.Each
WHENconditionisfollowedbyasetofoneormoreINTOclauses.
Foreachrowreturnedbythesubquery,eachWHENconditionisevaluatedto
beeithertrueorfalse.Iftrue,thefollowingsetofoneormoreINTOclausesis
executed.Iffalse,thesetofoneormoreINTOclausesisskipped,andthenext
WHENconditionisevaluated.
AnELSEclausemayoptionallybeincludedintheconditionalmultitable
INSERTstatement.Ifpresent,itmustdefineitsownsetofoneormoreINTO
clauses,andtheELSE/INTOclausesmustfollowallWHENconditions/INTO
clausecombinations.IfallWHENconditionsareskippedforanygivenrow,
thentheELSEclause’sINTOwillbeprocessed.Otherwise,itwillbeskippedfor
thatrow.
Eachrowreturnedbythesubqueryisprocessedaccordingtotheseruleswe
havejustreviewed.
Let’slookagainatourtableINVOICESandthearchivetable
INVOICES_ARCHIVED,inwhichwestoredinvoicerecordsthataremorethan
ayearold.SeeFigure13-1fortheINVOICEStable,andseeFigure13-2forthe
INVOICES_ARCHIVEDtable.
FIGURE13-2 DiagramfortheINVOICES_ARCHIVEDtable
Let’ssayourorganizationisengagedinamergerandwearetaskedwiththe
jobofintegratingdatafromanotherapplication.Thenewlyacquiredcompany
hasprovideduswiththetableWO_INV,asshowninFigure13-3.
FIGURE13-3 DiagramfortheWO_INVtable
WeneedtocreateanINSERTstatementthatwill
PulldatafromtheWO_INVtable
InsertWO_INV’sinvoiceinformationfromwithinthepastyearinto
ourINVOICEStable
InsertWO_INV’sinvoiceinformationthatismorethanayearold
intoourINVOICES_ARCHIVEDtable
It’saperfecttaskforaconditionalmultitableINSERTstatement,asfollows:
Inthisstatement,weseethefollowing:
Asubqueryonlines10through11.Notethesubqueryincludesa
columnDATE_SHIPPED.
Line2comparestheDATE_SHIPPEDvalueinaWHENcondition.
Ifline2evaluatestotrueforagivenrowfromthesubquery,the
INSERTstatementwilltakethatrow’sdataandinsertitintothe
INVOICES_ARCHIVEDtable,asspecifiedonline3.Thecolumnsin
theINVOICES_ARCHIVEDtablearespecifiedinlines3and4.
Line5definesthevaluesfromthesubquerythatwillbeinsertedif
theWHENclauseonline2istrue.Forexample,thesubquery’scolumn
INV_NO(line5)willbeinsertedintothetargettable’scolumn
INVOICE_ID(line3).
Line6isanELSEclausethatwillexecuteforeachrowthatdoesnot
satisfytheWHENconditioninline2.
Intheexamplewejustreviewed,therewasoneWHENconditionandone
ELSEcondition.Let’slookatanexamplewithmultipleWHENconditions.
Let’ssayyouhadthreearchivetables,namedINVOICES_THRU_2019,
INVOICES_THRU_2018,andINVOICES_THRU_2017,andwantedtoinsert
rowsfromtheincomingtableintoeacharchivedtablebasedontheyearofthe
DATE_SHIPPEDvalue.Notethateachtableisnotmutuallyexclusive;for
example,theINVOICES_THRU_2019tablewillcontaininvoicesfrom2019,
2018,and2017,aswellasearlier.Onerowreturnedbythesubquerymightbe
insertedintoallthreetables.
Toaccomplishthistask,youcouldusethefollowingINSERTstatement:
NoticethatthereisnokeywordFIRSTorALLinthisexample.Therefore,the
statementwilldefaulttoALL.SincetherearethreeWHENconditions,each
withanassociatedINTOclause,eachandeveryWHENconditionthatevaluates
totruewillexecute.Also,thisexampleomitstheELSEclause,soifanyrow
fromthesubquerydoesnotsatisfyaWHENcondition,thennoactionwillbe
takenforthatparticularrowreturnedbythesubquery.
AfteranyWHENcondition,youmayincludemorethanoneINTOclause.
Forexample,let’ssaywehaveatableINVOICES_CLOSEDthattakesany
invoicerowsthatshippedpriorto2018.Wemightmodifyourexamplelikethis:
NotethenewINTOclause,lines10through12.ThisINTOissubjecttothe
WHENconditioninline6.Inotherwords,ifDATE_SHIPPEDisintheyear
2018orbefore,theINSERTstatementwilladdthecandidaterowtoboththe
INVOICES_THRU_2018tableandtheINVOICES_CLOSEDtable.One
WHENconditionisthegatewaytobothINTOclauses.
WhatthisexampleshowsusisthatanyWHENconditioncanhavemultiple
INTOclausesthatfollowit.IftheWHENconditionevaluatestotrue,allofits
INTOclauseswillexecute.IftheWHENconditionevaluatestofalse,execution
willskipovertheINTOclausesandmoveondirectlytoeitherthenextWHEN
condition,anELSEifitispresent,orthenextrowinthesubquery.
TheINSERTALLwillevaluateeachandeveryWHENconditionandprocess
allINTOclausesforallWHENconditionsthatevaluatetotrue.Therefore,the
INSERTALLmayresultinasinglerowbeingaddedtomorethanonetable.
TheINSERTFIRSTwillevaluateeveryWHENconditionuntiloneofthem
evaluatestotrue.ItwillthenprocessthatWHENcondition’sINTOandskipthe
remainingWHENconditions.TheINSERTFIRSTwillprocessonlyzeroorone
WHENcondition;however,itmayalsoresultinasinglerowbeingaddedto
morethanonetable,butonlyifthefirsttrueWHENconditionhasmorethan
oneINTOclause.
TablealiasesinthesubqueryofamultitableINSERTarenotrecognizedin
therestoftheINSERT.Forexample,youcan’treferencethemfromwithina
WHENconditionorINTOstatement.Ifasubquery’scolumnreferencedepends
onatablealias,besuretouseacolumnaliasforthecolumnandthenreference
thecolumnaliasifyouwanttoreferencethatcolumnfromelsewherewithinthe
INSERTstatement—outsideofthesubquery.Forexample,seeFigure13-4.
FIGURE13-4 DiagramofthePOSITIONSandSALARY_CHARTtables
InaquerythatqueriesrowsfromthePOSITIONStableandconditionally
insertsthemintotheSALARY_CHARTtable,wecannotusethisquery:
Thisstatementwillnotwork.Hereistheresult:
Noticehowthesubqueryisaself-joinandusestablealiasestoidentifyeach
tableandcolumnreference.Thesetablealiasesareperfectlyfineinthesubquery
butarenotrecognizedbeyondthesubquery.Inotherwords,theattemptsto
referenceeachtablealiasfromwithintheWHENconditionorVALUESclause
areinvalid.
So,whatdowedo?Thesolutionistospecifyacolumnaliastoanycolumn
nameswithinthesubquerythatuseatablealiasandthenreferencethecolumn
aliasfromtherestoftheconditionalINSERTstatement,likewedointhe
followinglines5and6:
Notethatthisversionhasdonemorethanisrequired;itappliescolumn
aliasestoeachcolumninthesubqueryandthenreferencesthosecolumnaliases
fromtheWHENandVALUESclauses.Weonlyneededcolumnaliaseson
A.POSITIONandB.POSITIONinlines5and6,sowecanreferencethecolumn
aliasesinline4.Eitherway,thisversionoftheconditionalINSERTis
syntacticallycorrect.
YoucannotexecuteamultitableINSERTonaview;itcanbeusedonlywith
atable.
SequencegeneratorsdonotbehaveconsistentlyinamultitableINSERT
statement.Ifyoutrytouseasequencegeneratorwithinthesubquery,you’llget
asyntaxerror.IfyoutrytoincludeonewithintheexpressionlistoftheINTO
statement,youmayormaynotgetthefunctionalityyouwant—theNEXTVAL
functionwillnotadvanceasyoumightexpect.Thereasonisthatamultitable
insertistreatedasasingleSQLstatement.Therefore,ifyoureference
NEXTVALwithasequencegenerator,Oracle’sdocumentationwarnsthat
NEXTVALwillbeincrementedonceinaccordancewiththesequence
generator’sparametersandstaythatwayforthedurationofapassthroughthe
multitableinsert.Inotherwords,aconditionalINSERTwithasingleINTO,one
thatinvokesasinglesequencegeneratoroncewithaNEXTVAL,willincrement
thesequenceonceforeachrowreturnedbythesubquery—regardlessofwhether
theWHENconditionistrue.Considerthisexample:
Remember,atablealiasdefinedinasubqueryofamultitableINSERT
isnotrecognizedthroughouttherestoftheINSERTstatement.Also,ifa
multitableINSERTstatementfailsforanyreason,theentirestatementis
rolledback,andnorowsareinsertedinanyofthetables.
Thesequencegeneratorinline4willincrementforeachrowreturnedbythe
subquery,regardlessofwhethertheWHENconditionistrue.Forthisexample,
assumethesequencegeneratorhasjustbeencreatedandhasneverbeenused
andthatithasthedefaultsettingsofaninitialvalueof1andanincrementof1.
Giventhat,ifthesubqueryreturnstenrowsandif,forinstance,thefinalrow
alonecausestheWHENconditioninline2tobetrue,thentheonerowinserted
intotheINVOICES_ARCHIVEDtablewillbeassignedavalueof10forthe
INVOICE_IDcolumn.
Ifthisstatementcontainedadditionalcallstothesamesequencegeneratorin
additionalINTOclauses,theywouldnotcausethesequencegeneratorto
increment.Thesequencegeneratorincrementswitheachrowofthesubquery
returned—nomore,noless—regardlessofadditionalcallstoNEXTVAL.
Oracle’sdocumentationwarnsthat“youcannotspecifyasequenceinanypart
ofamultitableinsertstatement.”Theonlyplaceyou’llgetthesyntaxerrorisin
thesubquery,butknowthatattemptstoinvokeasequencegeneratorfromthe
WHENorINTOclauseoftheINSERTmayproduceundesirableresults.
UsingMultitableINSERTtoPerformaPivot
YoucanuseaconditionalmultitableINSERTstatementtotransformdatafroma
spreadsheetstructuretoarows-and-columnsstructure.Thissectiondescribesthe
technique.
First,let’sstartwiththefollowingdatalisting:
Thisisthesortofdatayoumightfindinatypicalspreadsheetdisplay.Let’s
saythisspreadsheethasbeenstoredinanexternaltable.Figure13-5showsthe
table’sstructure.
FIGURE13-5 DiagramoftheSHIP_CABIN_GRIDtable
Next,you’regiventhetaskofmovingthisdataintothetable,asshownin
Figure13-6.
FIGURE13-6 DiagramoftheSHIP_CABIN_STATISTICStable
Thisisn’tastraightforwardrow-for-rowinsertwithasubquery.Thisdata
mustbetransformedsothateachcolumnfromthespreadsheetistransformed
intoanindividualrowinthenewtable.
Thefollowingquerywillaccomplishthetask:
Notehoweachrowofthesubqueryisconsideredthreetimes.Foragivenrow
returnedbythesubquery,eachofthreecolumnsoftherowisindividually
considered.IfanyoneofthethreecolumnsOCEAN,BALCONY,and
NO_WINDOWisnotNULL,thenarowisinsertedintothetargettable.It’s
possiblethatsomeindividualrowsreturnedbythesubquerywillresultinthree
newrowsbeingaddedtothetargettableSHIP_CABIN_STATISTICS.
Let’stakealookattheresults:
Inthisexample,theconditionalmultitableINSERTtransformedincoming
datafromaspreadsheetsummarystyleintoarow-by-rowstructure,allwithina
singleSQLstatement.Inthisway,theconditionalmultitableINSERTstatement
“pivots”thedatabychangingcolumnsintorows.
NotethatthispivottechniqueisdifferentfromSQLoperationsthatusethe
keywordPIVOTorUNPIVOT.Whatwe’vedescribedhereisatechniquethat
usestheconditionalmultitableINSERTtopivotdata.ThekeywordPIVOT,
whilesomewhatsimilarinfunction,isaseparatefeature.NotethatOracle’s
publishedmaterial,includingdocumentspublishedinconnectionSQL
certificationexams,havedescribedthismultitableINSERTstyleasapivotand
makesadistinctionbetweenthispivotapproachandthePIVOTkeyword.
CERTIFICATIONOBJECTIVE13.02
MergeRowsintoaTable
TheMERGEstatementisaSQLDMLstatementthatcancombinethe
functionalityofINSERT,UPDATE,andDELETE,allintoasingleSQL
statement.Thereisn’tanythingyoucandowithMERGEthatyoucannotalready
dowithsomecombinationofthosethreeDMLstatements.However,ifit’s
possibletouseMERGEasanalternativetoexecutingtwoormoreDML
statements,thenMERGEispreferablesinceitcombinesmultipleDMLactions
intoasingleSQLstatement,resultinginasinglepassthroughthedatabase.In
otherwords,itwillperformmoreefficiently.Inmanysituations,suchas
applicationsbuiltwithdistributedarchitectures,MERGEmaybeaninvaluable
weaponinyourarsenaltobuildoptimallyefficientprofessionalapplications.
ThesyntaxofMERGEfollows:
Notethefollowing:
Line1INTOspecifiesthetargetintowhichyouareeitherinserting
orupdatingrows;itcanbeatableoranupdatableview.Thislineis
required.
Line2USINGidentifiesthesourceofthedata,whichcanbeatable,
view,orsubquery.Thislineisrequired.
Line3TheONconditionforMERGEbehavesessentiallylikea
WHEREclause.ItdetermineshowtocompareeachrowintheUSING
datasourcewitheachrowintheMERGEINTOdatatarget.TheON
conditioncanuseBooleanoperatorsandexpressionstoformcomplex
comparisons.Inpractice,theONconditionisoftenlimitedtocomparing
primarykeyvalues,butthisisnotrequired.Thislineisrequired.
Lines4through6Theselinesareconsideredthe“updateclause”
andidentifythelogicbywhichtheMERGEwillupdatetargetrows;it
cannotupdateacolumnintheONcondition.
Lines7through9Theselinesareconsideredthe“insertclause”and
identifythelogicbywhichtheMERGEwillinsertrowsintothetarget
table.
Asyoucansee,it’saninvolvedstatement.Let’slookatanexampleinaction.
Let’ssayyouareresponsibleforanapplicationthatusestheWWA_INVOICES
table(seeFigure13-7)andyouhavebeentaskedtobringindatafromanoutside
table,ONTARIO_ORDERS(seeFigure13-8).
FIGURE13-7 DiagramoftheWWA_INVOICEStable
FIGURE13-8 DiagramoftheONTARIO_ORDERStable
ThedatalistingfortheWWA_INVOICEStablefollows:
ThedatalistingfortheONTARIO_ORDERStablefollows:
Let’suseMERGEtobringthedatafromONTARIO_ORDERSintothe
WWA_INVOICEStable.
TheprecedingMERGEstatementincludesthefollowingfeatures:
Line1Herewespecifythatwearegoingtomergerowsintothe
tableWWA_INVOICES.Also,weassignatablealiasWWAtothetable
WWA_INVOICES.
Line2HerewespecifytheONTARIO_ORDERStableasthedata
sourceandgivethattableanaliasofONT.
Line3HerewedefinetheONcondition,indicatingthatthecolumns
CUST_POandPO_NUMarewherethecommoninformationexiststhat
will“join”therowslogicallyinordertoassociatethemwitheachother
forthemerge.
Lines4through5Theselinesarethe“updateclause.”
Lines6through10Theselinesarethe“insertclause.”
Line11ThislineistheWHEREclausefortheMERGE,filtering
outrowsfromtheUSINGdatasource—inthiscase,
ONTARIO_ORDERS.
TheresultofourMERGEistomergerowsfromONTARIO_ORDERSinto
WWA_INVOICES.IfwequerytheWWA_INVOICEStable,herearethe
results,aftertheMERGE:
Noticethefollowing:
WeupdatedtherowwhereCUST_POequalsWWA-001.
WeaddedtherowwhereCUST_POequalsWWA_017.
OurMERGEstatement’sWHEREclausecorrectlyignoredtherow
wherethePO_NUMwasNBC-201.
MERGEisausefulandefficientDMLstatement.Inthisexample,yousaw
howweperformedanINSERTandUPDATEstatementinasingleMERGE
statement.
OurexampledidnotmakeuseoftheDELETEfeatureofthe“updateclause”
withinMERGE.Butwecouldhaveincludedone,likethis:
TheUSINGclausecanbaseaMERGEonasubqueryaswellasa
tableorview.
Inthisexample,we’veaddedline6,whichcontainsthe“deleteclause”for
theMERGE.
Buttakenote:
The“deleteclause”affectsonlyrowsthatarearesultofthe
completed“updateclause”andremaininthetargettable—whichinthis
instanceisWWA_INVOICES.
Rowsaddedasaresultofthe“insertclause”areunaffectedbythe
“deleteclause.”
So,MERGErepresentsacombinationoftheUPDATEandINSERTDML
statementsand,toalesserandsomewhatlimitedextent,theDELETEstatement.
CERTIFICATIONSUMMARY
TheINSERTstatementcanbeaugmentedwithanumberofclausestointroduce
conditionallogicintoitsexecutionandtoadddatatomorethanonetablefrom
withinasingleINSERTstatement.Conditionallogiccanbeaddedwitha
WHENconditionandoptionallytheELSEkeyword.TheINSERTALLform
willtestincomingdataagainsteachWHENcondition,andtheINSERTFIRST
formwillstopatthefirstWHENconditionthatevaluatestotrue.Ineither
situation,theoptionalELSEclausecandefineaninsertthatwillexecuteifall
previousWHENconditionsfailed.ConditionalINSERTstatementsmaybeused
to“pivot”datafromcolumnsintorowsandbackagain.
TheMERGEstatementdoesnotdoanythingyoucannototherwisedowitha
seriesofotherDMLstatements,butitsadvantageisitspowerfulabilityto
performmultipleoperationsfromwithinasingleSQLstatementandthereforea
singleexecutionandsinglepassthroughthedatabase.
✓TWO-MINUTEDRILL
DescribetheFeaturesofMultitableINSERTs
Multitableinsertsareusefulforapplyingconditionallogictothe
databeingconsideredforinsertion.
Conditionallogiccanevaluateincomingrowsofdatainaseriesof
steps,usingseveralevaluationconditions,andcanofferalternative
strategiesforaddingdatatothedatabase,allinasingleSQLstatement.
MultitableINSERTstatementsofferflexibilityandperformance
efficiencyoverthealternativeapproachesofusingmultipleSQL
statements.
MultitableINSERTstatementsmayuseconditionaloperationssuch
astheWHENconditionandtheELSEclause.
AWHENconditioncanbeusedtoevaluateincomingdataand
determinewhetheritshouldbeinsertedintothedatabase,andifyes,
whichtableandwhichcolumnsaretobeinserted.
TheELSEclauseisalastalternativechoicethatwillexecuteifno
WHENconditionevaluatedtotrue.
BothWHENandELSEareassociatedwiththeirownunique
INSERTstatementdirectives;dependingonwhichconditionsapply,the
appropriateINSERTstatementdirectiveswillexecute.
EachconditioncanINSERTdataindifferentwaysintodifferent
tables.
TheINSERTFIRSTstatementtestseachWHENconditionand
executestheassociatedINSERTstatementdirectiveswiththefirst
WHENconditionthatevaluatestotrue.
TheINSERTALLstatementexecutesalltheWHENconditionsthat
evaluatetotrue.
TheELSEclauseexecutesforeithertheINSERTFIRSTor
INSERTALLstatementwhennoneoftheWHENconditionshas
executed.
ThesubqueryofamultitableINSERTdeterminesthedatathatwill
beconsideredintheinsertlogic;itcanbeacomplexqueryandcan
includejoins,GROUPBYclauses,setoperators,andothercomplex
logic.
MergeRowsinaTable
TheMERGEstatementisoneoftheSQLDMLstatements,
alongsideSELECT,INSERT,UPDATE,andDELETE.
MERGEreplicatessomeofthefunctionalityfoundinINSERT,
UPDATE,andDELETEandcombinesitallintoasinglestatementthat
executeswithasinglepassthroughthedatabase.
MERGEdoesn’tdoanythingnewthatyoucannotalreadydowith
existingDMLstatements,butitdoesthemmoreefficientlyin
combination.
TheMERGEstatementincludesan“updateclause”andan“insert
clause.”
TheWHENMATCHEDTHENUPDATEkeywordsformthe
“updateclause.”
TheWHENNOTMATCHEDTHENINSERTkeywordsformthe
“insertclause.”
TheDELETEclauseoftheMERGEstatementonlydeletesrows
thatwerefirstupdatedwiththe“updateclause”andremainaftera
successfulupdate;theymustalsomeettheWHEREconditionofthe
“deleteclause.”
SELFTEST
Thefollowingquestionswillhelpyoumeasureyourunderstandingofthe
materialpresentedinthischapter.Chooseonecorrectanswerforeachquestion
unlessotherwisedirected.
DescribetheFeaturesofMultitableINSERTs
1.WhatcananINSERTstatementdo?(Choosetwo.)
A.Addrowsintomorethanonetable
B.Adddataintomorethanonecolumninatable
C.Deleterowsbyoverwritingthem
D.Jointablestogether
2.AmultitableINSERTstatement:
A.Canaccomplishtasksthatcannototherwisebedoneinany
combinationofSQLstatements
B.WillcreateanytablesinwhichitattemptstoINSERTbutthatdo
notyetexist
C.Canuseconditionallogic
D.Iscapableofinsertingrowsintononupdatableviews
3.ReviewthefollowingdiagramsoftheSPAREStable:
AlsoexaminethediagramsofthetablesPORT_INVENTORY,
STORE_INVENTORY,andSHIP_INVENTORY,shownhere.
NowconsiderthefollowingSQLstatement:
RegardingthisSQLstatement,whichofthefollowingstatementsis
true?
A.ThestatementwillfailbecausethereisnoELSEclause.
B.ThestatementwillfailbecauseitismissingaWHENcondition.
C.ThestatementwilladdarowreturnedfromtheSPAREStableto
theSHIP_INVENTORYtableonlyiftheWHENconditiononline2
evaluatestotrue.
D.ThestatementwilladdeveryrowreturnedfromtheSPARES
tabletotheSHIP_INVENTORYtable.
4.ReviewtheSQLstatementintheprecedingquestion.Ifoneofthe
INTOclausesexecutedonatableandresultedinaconstraintviolationon
thattable,whatwouldresult?
A.Therowwouldnotbeinserted,andtheINSERTstatement
wouldskiptothenextrowreturnedbythesubqueryandperform
anotherpassthroughtheWHENconditions.
B.Therowwouldnotbeinserted,andtheINSERTstatementwould
stop.Noadditionalrowswouldbereturnedbythesubqueryor
processed,butrowsthathavealreadybeenprocessedareunaffected.
C.Therowwouldnotbeinserted,theINSERTstatementwould
stop,andallrowsaffectedbytheINSERTstatementwouldberolled
back,asiftheINSERTstatementhadneverbeenexecuted.
D.Noneoftheabove.
5.Reviewthediagramsinquestion3,andconsiderthefollowingSQL
statement:
Whichoneofthefollowinganswerscorrectlyidentifiesdatathat,if
presentintheSPAREStable,willbeinsertedbythisconditionalINSERT
statementintothetable—ortables—identifiedbytheanswer?
A.PART_NO=123;PART_NAME=‘BAH-OPS,’inboth
STORE_INVENTORYandPORT_INVENTORY
B.PART_NO=401;PART_NAME=‘PAN-OPS,’inboth
SHIP_INVENTORYandPORT_INVENTORY
C.PART_NO=170;PART_NAME=‘TRA-OPS,’in
STORE_INVENTORY
D.PART_NO=4;PART_NAME=‘PAN-OPS,’inboth
STORE_INVENTORYandSHIP_INVENTORY
6.Reviewthediagramsinquestion3,andexaminethefollowing
statement:
WhichofthefollowingstatementsistrueforthisSQLstatement?
A.IfthefirstWHENconditioninline2istrue,theINTOclausein
line3andline4willbeexecuted,afterwhichprocessingwillskipto
thenextrowreturnedbythesubquery.
B.IfthefirstWHENconditioninline2istrue,theWHEN
conditioninline7willnotbeevaluated.
C.NomatterwhichWHENconditionistrue,theINTOclausein
line5willbeexecutedregardless.
D.RegardlessofwhetherthefirstWHENconditionistrue,the
secondWHENconditionwillbeevaluated.
MergeRowsinaTable
7.TheMERGEstatementincludesaUSINGclause.Whichofthe
followingstatementsisnottrueoftheUSINGclause?
A.Itcanbeusedtospecifyasubquery.
B.ThedataitidentifiesremainsunchangedaftertheMERGE
statementexecutes.
C.TheUSINGclauseisoptional.
D.Itcanbeusedtospecifyaninlineview.
8.Seethediagramsinquestion3.Youwanttomergerowsfromthe
PORT_INVENTORYtableintotheSHIP_INVENTORYtable.Youstart
withthefollowingSQLstatement:
WhatwillthisSQLstatementdo?
A.ItwillfailwithasyntaxerrorbecauseyoumusthaveanELSE
clause.
B.Itwillfailwithasyntaxerrorbecauseyoucannotreferencethe
targettable(SHIP_INVENTORY)intheWHEREclauseinline8.
C.ItwilladdrowsfromPORT_INVENTORYto
SHIP_INVENTORYthatdonotalreadyexistinSHIP_INVENTORY,
limitedtoLAST_ORDERvaluesfromtheyear2019.
D.ItwilladdrowsfromPORT_INVENTORYto
SHIP_INVENTORYthatdonotalreadyexistinSHIP_INVENTORY,
regardlessofthevalueforLAST_ORDER.
9.ExaminetheSQLsyntaxinquestion8.Whichofthefollowingtwo
alternativesforline3aresyntacticallycorrect?
A.Onlyoption1
B.Onlyoption2
C.Bothoption1andoption2
D.Neitheroption1noroption2
10.Whichofthefollowingstatementsisfalse?
A.Itispossibletomergeintotwoormoretables.
B.Itispossibletomergeintoaview.
C.TheUSINGclausecanreferencetwoormoretables.
D.Youcannotperformanupdatetoacolumnthatisreferencedin
theONclause.
SELFTESTANSWERS
DescribetheFeaturesofMultitableINSERTs
1. AandB.INSERTstatementscanaddrowstomorethanonetable
usingconditionalandunconditionallogic.INSERTstatementscanalsoadd
datatomorethanonecolumninanygiventable.
CandDareincorrect.INSERTcannotoverwritedata;itaddsnewrows
toatable.INSERTdoesnotperformjoins.
2. C.MultitableINSERTstatementscanuseconditionallogic,with
statementssuchasWHENandELSE.
A,B,andDareincorrect.MultitableINSERTSdonotdoanythingyou
couldn’totherwisedowithoneormoreSQLstatements.Theiradvantageis
thattheycanaccomplishcomplexSQLtasksinasinglepassthatmight
otherwiserequiremultiplepassesthroughthedatabase,thusyielding
performanceadvantages.Andnothingcanaddrowsintoanonupdatable
view—ifit’snotupdatable,it’snotupdatable.
3. C.TheWHENconditioninline2determineswhethertheINTO
clausesinlines3,4,5,and6willexecute.
A,B,andDareincorrect.TheELSEclauseisnotrequired.Noparticular
WHENconditionisrequired.TheINTOclauseforSHIP_INVENTORYis
subjecttotheWHENconditioninline2.
4. C.Theentirestatementfails,andallinsertedrowsarerolledback.It
isasifthestatementhadneverbeenexecuted.Hadthisstatementincluded
anycallstoasequencegeneratoranditsNEXTVALpseudocolumnwould
haveadvancedthecountinthegenerator,thateffectwouldremain
unchanged.However,thisexampledoesnotincludeanysequence
generators,sothatparticularexceptiondoesnotapply.
A,B,andDareincorrect.
5. C.ThePART_NOof170hasalengthof3,andthatislongerthan2,
sotheWHEREclauseinline13isfoundtobetrue,andtherowwillbe
evaluatedbytherestoftheINSERTFIRSTstatement.Next,the
PART_NAMEofPAN-OPSwillcausethefirstWHENconditiontobetrue,
andsincethisisanINSERTFIRSTstatement,nootherWHENcondition
willbeconsidered.
A,B,andDareincorrect.Theseanswersresultfromvarious
interpretationsoftheWHENconditionsandELSE.InanINSERTFIRST
statement,thefirstWHENconditionthatevaluatestotrueistheonly
conditionthatisexecuted.Allothersareignored.IfnoWHENisfoundto
betrue,thentheoptionalELSEclausewillbeprocessed.
6. D.BothWHENconditionswillbeevaluatedbecausethe
conditionalINSERTisanINSERTALLstatement.
A,B,andCareincorrect.IfthefirstWHENconditionistrue,bothINTO
clausesthatfollowitwillbeexecuted—thatincludestheINTOonline5
throughline6.WhetherthefirstWHENconditionistrueorfalse,the
secondwillalsobeevaluatedsincethisisanINSERTALLstatement.The
INTOinline5throughline6willbeevaluatedonlyifthefirstWHEN
conditionistrue.
MergeRowsinaTable
7. C.TheUSINGclauseisnotoptional;itisrequiredintheMERGE
statement.
A,B,andDareincorrect.USINGcanidentifyatable,view,orsubquery.
Aninlineviewisalsoacceptable.Itidentifiesthesourceofdatatobe
merged;thesourcedataremainsunchangedaftertheMERGEstatementis
executed.
8. B.ItwillfailbecausetheWHEREclausereferencessomethingthat
isnotinthesourcetable.TheWHEREclauseisanextensionofUSING,
whichspecifiesthetargettable.TheAtablealiasreferenceismeaningless
andwillfail.
A,C,andDareincorrect.ThereisnoELSEclauseinMERGE,soitis
notonlynotrequired,itisnotaccepted.
9. C.Bothoptionsareacceptable.TheONconditioncanbeany
comparisonofexpressions,anditcanincludeBooleanoperators.
A,B,andDareincorrect.
10. A.YoucannotMERGEintotwoormoretables.Onlyoneispermitted.
B,C,andDareincorrect.YoucanMERGEintoaviewprovidedthe
viewisupdateable.TheUSINGclausecanreferencetwoormoretablesby
wayofajoinorsubquery.Youcannotchangethevaluesofajoincriteria
duringthejoin,soyoucannotupdatecolumnsthatarereferencedinanON
clause.
14
ControllingUserAccess
T
CERTIFICATIONOBJECTIVES
14.01DifferentiateSystemPrivilegesfromObjectPrivileges
14.02GrantPrivilegesonTablesandonaUser
14.03DistinguishBetweenPrivilegesandRoles
✓Two-MinuteDrill
Q&ASelfTest
hischapterexploresthesubjectofuseraccessandtheprivilegesassociated
withperformingactionsinthedatabase.Everyactionperformedbyanyuser
accountrequiresacorrespondingprivilegeorsetofprivilegestoperformthat
action.Therearetwocategoriesofprivileges.Systemprivilegesarerequiredto
performataskinthedatabase;objectprivilegesarerequiredtousethosesystem
privilegesonanygivendatabaseobjectinparticular.Privilegesmaybegranted
toauseraccountortoanotherdatabaseobjectcalledarole.Arole,inturn,can
begrantedtoauseraccount,whicheffectivelygrantsthesetofprivileges
collectedwithintherole.Oncegranted,privilegesandrolescanlaterberevoked.
Together,privilegesandrolesarethemechanismformanagingandcontrolling
accesstothedatabasebyuseraccounts.Thischapterlooksathowtocreateand
manageprivilegesandroles.
Awordofwarningaboutthesamplecodecontainedinthischapter:someof
ithastheabilitytochangeyourdatabasepermanentlywithresultsthatmaybe
undesirable.SomeofourcodesampleswilllookatSQLcodethatusesthe
SYSTEMuseraccount,animportantaccountthatshouldbecontrolledby
experienceddatabaseadministratorsinanyproductiondatabase.Youshould
alwayscheckwithyourdatabaseadministrator(DBA)beforetryinganycode
samplesfromanybook,butthischapterinparticularincludescodethatyou
shouldnotexecuteinaprofessionalinstallationwithoutfirstcheckingwithyour
DBA.
CERTIFICATIONOBJECTIVE14.01
DifferentiateSystemPrivilegesfromObjectPrivileges
Throughoutthisbook,we’velookedathowauseraccountcanuseSQL
statementstocreateanduseavarietyofdatabaseobjects.However,beforeany
useraccountcanexecuteaSQLstatement,itmustbegrantedtheprivilegeto
executethatSQLstatement.Furthermore,onceadatabaseobjecthasbeen
created,anyuseraccountthatwillusethedatabaseobjectmustfirstbegranted
privilegestodoso.
Therearethreegeneralcategoriesofprivileges,asdescribedinTable14-1.
TABLE14-1 TypesofPrivileges
We’llrevieweachoftheitemslistedinTable14-1inthischapter.
SystemPrivileges
Systemprivilegesaretherighttoperformsometaskinthedatabase.For
example,tologintothedatabase,auseraccountisgrantedthesystemprivilege
CREATESESSION.Tocreateatable,auseraccountmustbegrantedthesystem
privilegeCREATETABLE.
Therearemorethan100differentsystemprivilegesthatcanbegrantedtoa
useraccount.Table14-2listssomeofthesystemprivilegesthatarerequiredto
performthetaskswe’vediscussedinthisbook.
TABLE14-2 SomeSystemPrivileges
Systemprivilegesdifferfromobjectprivilegesinthatsystemprivilegesare
whatauseraccountmusthavetocreatedatabaseobjects,amongotherthings.
Then,onceadatabaseobjecthasbeencreated,objectprivilegesonthatdatabase
objectcanbegrantedtootherusers.
Forexample,therighttoexecutetheSQLstatementCREATETABLEand
createanewdatabasetableisasystemprivilege.Buttheabilitytochangerows
ofdatain,forexample,atablecalledBENEFITSownedbyauseraccount
namedEUNICEisanobjectprivilege.Inotherwords,anobjectprivilegeisthe
righttodosomethingtoaparticularobject.
Asananalogy,considertheconceptofadriver’slicense.Adriver’slicenseis
sortoflikeasystemprivilege;it’stherighttodriveacarinageneralsense.
Onceyouhaveadriver’slicense,ifyougetacar,youcandriveit.Butyoudon’t
havetherighttodriveanyone’scarinparticularunlesstheownerspecifically
authorizesyoutodoso.
Therighttodrivesomeoneelse’scarislikeanobjectprivilege.Youneed
bothoftheseprivilegesinordertodriveacarandtobeinfullcompliancewith
thelaw.Thesameistrueinthedatabase:youneedsystemprivilegestoperform
particulartasks,andyouneedobjectprivilegestoperformthosetasksonan
objectinparticular.
Let’slookatsomeofthesyntaxforgrantingprivileges.Notethatforsomeof
theupcomingexamples,we’llusetheSQL*PlustoolandsomeSQL*Plus
commands.TheseSQL*Pluscommandsdonotrequirethesemicolon
terminationcharacterthatisrequiredinSQLstatements.
We’llusetheSQL*PluscommandCONNECTtologintoanotheruser
account.YoucanalsousetheSQL*PluscommandSHOWUSERtoconfirm
whichaccountiscurrentlyactiveinthesession.SQL*Pluscommandsare
helpfultouseinyourSQLsessions.Buttheyarenotontheexam.
Prerequisites
BeforewegetstartedwithGRANTandREVOKEstatements,let’sreviewsome
supportingstatementsthataren’tspecificallyincludedintheexamobjectivesbut
areusefulfordemonstratingsystemprivileges,objectprivileges,roles,andtheir
capabilities.
CREATE,ALTER,andDROPUSER
Let’slookathowtocreateauseraccount.AnySQLuserwiththeCREATE
USERsystemprivilegemayexecutetheCREATEUSERstatement,whose
syntaxlookslikethis:
CREATEUSERusernameIDENTIFIEDBYpassword;
Inthisstatement,usernameisanameyouspecifyaccordingtotherulesfor
namingdatabaseobjects.Thepasswordfollowsthesamerules.(Notethat
passwordsarecasesensitivebydefaultstartingwithOracle11g.)
Forexample,thisstatementwillcreateausernameJOANwiththepassword
DEMERY:
CREATEUSERJOANIDENTIFIEDBYDEMERY;
YoucanusetheALTERUSERstatementtochangethepassword,likethis:
ALTERUSERJOANIDENTIFIEDBYHAWAII;
Finally,youcanremoveauserfromthedatabaseusingtheDROPUSER
statement,likethis:
DROPUSERusername;
Ifauseraccountownsanydatabaseobjects,theprecedingstatementwon’t
work,andyou’llneedtousethis:
DROPUSERusernameCASCADE;
TheCASCADEoptiondirectsSQLtodroptheuseraccountandallofthe
objectsitowns.
Onceauserobjecthasbeencreated,itcanbegrantedprivileges,asyou’llsee
inanupcomingsection.
CONNECT
TheCONNECTstatementisnotaSQLstatementbutaSQL*Plusenhancement
youcanusewithintheOracleSQL*Plustool.Onceyou’vestartedSQL*Plus,
youcanuseCONNECTtologinorswitchloginsessionsfromoneuseraccount
toanother.If,forexample,youareusingtheSQL*Plustoolandhaveloggedin
totheEFCODDaccountandcreatedtheuseraccountJOAN,youcanloginto
theJOANaccountdirectlyfromEFCODDwiththisstatement:
CONNECTJOAN/HAWAII
ThisassumestheuseraccountJOANisstillusingthepasswordHAWAII.It
alsoassumesthatJOANhasbeengrantedtheminimumsystemprivilegestolog
in,suchasCREATESESSION.
Again,asemicolonterminationcharacterisnotrequiredinSQL*Plus
statements.Itisacceptedbutnotrequired.Thesemicolonterminationcharacter
isrequiredinSQLstatementsbutisoptionalinSQL*Plusstatements.
Tablespaces
Inthecourseofsettingupanewuseraccount,thetopicoftablespacesmustbe
addressed.However,thetopicoftablespacesgoesbeyondourscopeandisnot
includedintheexamobjectives,sowe’llshowthesimplewaytoaddressthe
tablespacerequirement,asfollows:
ThiswouldprobablynotbesomethingthatyourtypicalproductionDBA
woulddo.Tablespacesarecontrolledbydatabaseadministrators.AtypicalDBA
generallycreatesuniquelynamedtablespacesandcarefullyallocatesspace
quotastothem.We,however,aren’tconcernedwithanyofthatforthisbookor
fortheexam.Soforus,theprecedingstatementiswhatwe’llinclude.Ifyou
wanttolearnmore,weencourageyoutocheckoutanyoftheoutstandingbooks
fromOraclePressonthetopicofdatabaseadministration.Inthemeantime,if
you’reworkingonyourowntestsystemonyourownpersonalmachine,this
particularstatementthatgrantsUNLIMITEDTABLESPACEismorethan
adequateforourpurposesgoingforward.Ifyou’reusingtheseatwork,check
withyourDBAbeforetryinganyofthecodesamplesinthischapter.
GRANTandREVOKE
Nowlet’sgetdowntobusiness.SystemprivilegesaregrantedwiththeGRANT
statement.Here’sanexampleofaSQLsessionthatlogsintotheOracle
SYSTEMaccount,createsanewuser,andgrantsthenewuseraccountsome
initialsystemprivilegesusingthreeGRANTstatements(linenumbersadded):
Inthesestatements,hereiswhatwearedoing:
Line1WeestablishausersessionwiththeuseraccountSYSTEM,
withapasswordofMANAGER.TheSYSTEMaccountisinstalledwith
everyOracledatabase,andtheDBAinstallingOracleassignsthe
password.(Warning:Donottrythisonaproductionsystem.Noself-
respectingproductionsystemshouldhaveaSYSTEMaccountpassword
settoavalueofMANAGERanyway,butthepointisthatifyouhave
installedyourownversionoftheOracledatabaseonyourownlocal
machineanditisnotusedforproductionwork,thenyoucantrythis,but
ifyou’retryingthingsoutwithinasystematyourworkplaceor
somewherecomparable,thenbesuretocheckwithyourdatabase
administratorbeforetryingthis.)
Line2WecreateanewuseraccountcalledHAROLD,withthe
passwordLLOYD.
Line3WeusetheSQLstatementGRANTtogivetheCREATE
SESSIONprivilegetouserHAROLD.Thisisaminimumrequirement
forustobeabletologintothedatabasewiththeHAROLDuser
account;withoutthisGRANTstatement,wecouldn’tsuccessfullylogin
withtheuseraccountHAROLD.
Line4ThisisonewaytoensurethatHAROLDcancreateobjects.
Seeourearlierdiscussionabouttablespacesintheprevioussection.
Line5UsingGRANT,wegivethesystemprivilegeCREATE
TABLEtouseraccountHAROLD.
SeeFigure14-1fortheresultsofthesestatementsintheSQL*Pluswindow.
FIGURE14-1 SQL*Plussession:GRANTstatements
Nowlet’slogintoHAROLDandtryoutwhatwe’vedone.Forthat,we’lltry
thefollowingSQLstatements:
SeeFigure14-2fortheresults.Notethatwearen’tabletocreatethesequence
becausewehaven’tbeengrantedsufficientprivilegestodoso.Forthat,we’ll
needtologbackintotheSYSTEMaccountandgrantthesystemprivilege
CREATESEQUENCEtoHAROLD.Oncethathasbeenaccomplished,wecan
logbackintoHAROLDandcreatethesequence(seeFigure14-3).
FIGURE14-2 SQL*Plussession:testingsystemprivileges
FIGURE14-3 SQL*Plussession:creatingthesequence
Intheseexamples,wehavebeenloggedintotheSYSTEMaccounttogrant
theseprivileges,butanyqualifiedDBAaccountwilldoandispreferableinany
seriousinstallationwithmultipleOracleusers.Insuchasituation,thelesstimea
developerorDBAspendsintheSYSTEMaccount—ortheotherrestricted
defaultDBAaccountsintheOracledatabasesuchasSYS—thelesslikelya
mistakewillaccidentallycausesomeseriousdamagetothedatabase.
ThebasicsyntaxfortheGRANTstatementissimple.
Here,privilegeisoneoftheseveraldozensofsystemprivilegesthatare
alreadydefinedinthedatabase(seetheOracleDatabaseSQLLanguage
ReferenceManualforacompletelist).Multipleprivilegescanbegrantedatonce
byseparatingeachadditionalprivilegewithacomma,asinGRANTprivilege,
privilege.(We’lldiscussoptioninanupcomingsection.)
ThebasicsyntaxforREVOKEiscomparable.
NotethatyougrantTOandyourevokeFROM.
Onceasystemprivilegeisrevokedfromauser,theeffectisimmediate.
However,anyactionstakenpriortotherevocationstand.Inotherwords,ifa
useraccounthasbeengrantedthesystemprivilegeCREATETABLEandthen
createssometablesbutthenhastheCREATETABLEsystemprivilegerevoked,
thecreatedtablesalreadyinexistenceremaininplace.Theydonotdisappear.
ButtheowningusermaynotcreateadditionaltableswhiletheCREATETABLE
systemprivilegeisrevoked.
We’velookedatafewsystemprivileges,andwe’vesaidthattheyare
somewhatlikeadriver’slicense.Nowlet’sextendtheanalogyalittlebit:
imaginewhatwouldhappenifyoucouldgetauniversaldriver’slicensethat
carriedwithittheabilitytodriveanyone’scarlegallywithoutthecar’sowner
expresspermission.SuchaconceptexistswithintheOracledatabase,andit’s
embodiedinthekeywordANY.Let’slookatthatnext.
ANY
SomesystemprivilegesincludethekeywordANYinthetitle.Forexample,
thereisasystemprivilegeCREATEANYTABLE,whichistheabilitytocreate
atableinanyuseraccountanywhereinthedatabase.Let’slookatasample
sessionthatinvolvesthisprivilege.
TheresultoftheprecedingSQLstatementsisthattwouseraccountswillbe
created.Also,twotableswillbecreated,withonetablecalledMOVIESand
anothertablecalledTVSHOWS.Bothtableswillexistintheuseraccount
LAUREL.ThefirsttablewascreatedbyLAUREL,butthesecondtable,
TVSHOWS,wascreatedbyuseraccountHARDYandwascreatedasatable
thatisownedbyLAUREL.TheuseraccountHARDYwillcontainnotables.
Theofficial“owner”ofbothtablesisLAUREL,asthedatadictionaryconfirms.
WhenasystemprivilegeincludesthekeywordANYinitstitle,itmeansthat
theprivilegewillauthorizeausertoperformthetaskasthoughtheywereany
useraccount.Inthisexample,userHARDYwasabletocreateatableandplace
itintheLAURELaccount,atasktypicallyreservedonlyforuserLAUREL.
However,sinceuserHARDYhasthesystemprivilegeCREATEANYTABLE,
thenHARDYcancreateanytableinanyuseraccount.
ADMINOPTION
InaprevioussectionwesaidwewouldlookattheoptionintheGRANT
statement’ssyntaxweexamined.Hereitis:theoptionisanadditionalclausethat
maybeincludedwiththeGRANTstatement,asfollows:
WhenanysystemprivilegeisgrantedwiththeWITHADMINOPTION
option,thentherecipientreceivesthesystemprivilegeitself,alongwiththeright
tograntthesystemprivilegetoanotheruser(seeFigure14-4).
FIGURE14-4 GRANTversusGRANTWITHADMINOPTION
TheREVOKEstatementdoesnotusetheWITHADMINOPTIONclause.
Wheneverasystemprivilegeisrevoked,theentiresystemprivilegeisrevoked.
Ifauser—let’scallitthefirstuser—grantsasystemprivilegetoasecond
userWITHADMINOPTIONandtheseconduserusestheadminoptionto
grantthatsamesystemprivilegetoathirduser,thenthethirduserretainsthe
privilegeuntilitisexplicitlyrevokedfromthethirduser.Inotherwords,once
theseconduserhasgrantedthethirduserwiththesystemprivilege,itstayswith
thefirstuser,evenifthefirstuser—oranyotherqualifieduser—revokesthe
systemprivilegefromtheseconduser.Ifthathappens,thethirduserstillhasthe
systemprivilege.Theonlywaythethirduserwilllosethesystemprivilegeisif
anyqualifieduserrevokesthesystemprivilegeexplicitlyfromthethirduser
withaREVOKEstatement.Inotherwords,theREVOKEstatementforsystem
privilegesdoesnot“cascade.”Itappliesonlytotheusertowhomtherevocation
isapplied.
ALLPRIVILEGES
Asanalternativetograntingspecificsystemprivileges,aqualifieduseraccount,
suchasSYSTEMorsomeotherDBAqualifiedaccount,canissuethefollowing
statement:
Thisstatementhastheeffectofgrantingallsystemprivilegestotheuser.The
WITHADMINOPTIONclausemaybeusedwiththisaswell.
Needlesstosay,thisshouldbedonewithgreatcaution,ifatall.Itisnot
easilyreversible.Inotherwords,thefollowingisnotanexactcounterpart:
Thisstatementwillreverseallsystemprivilegesgrantedtotheuser,assuming
thatallsystemprivilegeshavebeengrantedtotheuser.Ifnot,anerrormessage
willresult.
PUBLIC
ThePUBLICaccountisabuilt-inuseraccountintheOracledatabasethat
representsallusers.AnyobjectsownedbyPUBLICaretreatedasthoughthey
areownedbyalltheusersinthedatabase,presentandfuture.
TheGRANTstatementwillworkwiththekeywordPUBLICintheplaceofa
useraccountname.Here’sanexample:
ThisstatementgrantstheCREATEANYTABLEprivilegetoeveryuserin
thedatabase.TheCREATEANYTABLEprivilegegiveseveryusertheability
tocreateanytableinanyotheruseraccount.Inotherwords,it’smasshysteria—
orsomethinglikeit.Mindyou,we’renotrecommendingyoudothis,butit’s
syntacticallypossible,andyouneedtobeawareofit.Whilethissortofan
exampleisunlikely,grantingtoPUBLICmaybeusefulwithaselectednumber
ofobjectprivileges,whichwe’lldiscussabitlater.
Notethatifyoucometoyoursensesanddecidetorevokeasystemprivilege
fromPUBLIC,youcandosowithoutrevokinganyothersystemprivileges.In
otherwords,considerthisstatement:
REVOKECREATEANYTABLEFROMPUBLIC;
Notethatifyouwanttograntallprivileges,youusethekeywordsALL
PRIVILEGES.Butifyouwanttograntcertainprivilegestoallusers,
youdonotusethekeywordALL.Instead,yougranttoPUBLIC.
ThisstatementwillreversetheGRANT…TOPUBLICthatweissuedafew
paragraphsearlierandthankfullywillnotrevokeanyindividuallygranted
CREATEANYTABLEsystemprivilegesheldbyanyuseraccounts.Itwill
revokeonlytheGRANTtoPUBLIC.
Ifyou’reeventhinkingaboutusingGRANTALLPRIVILEGESTOPUBLIC
WITHADMINOPTION,youcanputthatthoughtoutofyourmindrightthis
second.
CERTIFICATIONOBJECTIVE14.02
GrantPrivilegesonTablesandonaUser
AnyuserwiththesystemprivilegeCREATETABLEcancreateatable.The
table,oncecreated,isownedbytheuserwhocreatedit.Theownerdoesnot
requireanyexplicitlygrantedprivilegesonthetable.Thetableownercanuse
DMLtoaddrows,changedatainthetable,querythedatainthetable,and
removerowsfromthetable.Butotherusersdonothavethatprivilege
automatically.Otherusersmusthaveexplicitlygrantedprivilegesontheobject,
which,inthiscase,isatable.
(Note:Theexception,ofcourse,isthoseuserswhohavethesystem
privilegesthatallowthemtorunanyDMLstatementsonanytableinthe
database,regardlessofwhoownsit.Thosesystemprivileges,aswesawinthe
previoussection,includeSELECTANYTABLE,INSERTANYTABLE,
UPDATEANYTABLE,andDELETEANYTABLE.)
Anyuserwhoownsatable—oranyotherdatabaseobject—maygrantobject
privilegesontheirdatabaseobjecttootherusersinthedatabase.
ObjectprivilegesexistforallDMLstatements—SELECT,INSERT,
UPDATE,andDELETE—aswellasanyDDLstatementthatisrelevanttoan
existingobject,suchasALTER,forexample.Notethatthereisnoseparateset
ofobjectprivilegesfortheMERGEstatement.
ObjectprivilegesonatableincludealltheDMLstatementsthatcanbe
executedagainstatable.Forexample,ifauseraccountLISAhasthesystem
privilegeCREATETABLE,thenLISAcancreateatable.IfLISAtakes
advantageofthissystemprivilegeandcreatesatableWEBINARS,thenLISA
canaccessthenewtable,butotherusersarenotautomaticallyabletoseethe
table(unless,aswestatedearlier,thoseuserspossessoneoftheANYsystem
privileges,suchasSELECTANYTABLE).Toensurethatotheruseraccounts
canexecuteSQLstatementsonthetable,useraccountLISAwillhavetogrant
objectprivilegesonWEBINARStootherusers.
SeeFigure14-5foraSQL*PlussessioninwhichweconnecttotheSYSTEM
account,wherewecreatetwouseraccounts,LISAandHENRY.WegiveLISA
sufficientprivilegestoconnect(CREATESESSION)andcreatetables.Note
howwecombinedmultiplesystemprivilegesinasingleGRANTstatement.
Also,wegiveHENRYsufficientprivilegestocreateasession—butnothing
more.
FIGURE14-5 Creating,granting,andtestingobjectprivileges—part1
WecontinueinFigure14-6,whereweconnecttotheLISAaccountandcreate
atable,adddatatoit,andthengrantprivilegesonthetabletoHENRY.Thenwe
connecttoHENRY,wherewecanissueSELECTandUPDATEstatementsbut
notINSERT—thatparticularprivilegewasn’tgrantedtoHENRY.
FIGURE14-6 Creating,granting,andtestingobjectprivileges—part2
TakeanotherlookatFigure14-6,andnotethemomentthattheGRANT
statementisissued.RememberthatanyDDLstatementcarrieswithitan
implicitcommitevent.Inotherwords,theGRANTstatementhastheeffectof
makingtheresultsoftheINSERTstatementpermanentinthedatabase.Once
thatGRANThasexecuted,theoptiontorollbacktheINSERTstatementwith
ROLLBACKisnolongeravailable.
SchemaPrefixes
NoteinFigure14-6thatwhenHENRYreferencesatableownedbyLISA,
HENRYmustusetheschemaprefixtomakethereference.Inotherwords,
HENRYcouldnotissueaSELECTstatementlikethis:
ASYNONYMisanobjectinthedatabasethatisanalternativenamefora
databaseobject.APUBLICSYNONYMisaSYNONYMthatisownedbythe
PUBLICuseraccount,whichisanautomaticallycreateduseraccountthatis
maintainedbytheOracledatabase.ThePUBLICuserisn’tintendedtobean
accountintowhichyoulogintogetaccess.Instead,PUBLICisamechanismby
whichyoucancreategloballyownedobjects.Specifically,anythingthatis
ownedbyPUBLICisautomaticallyownedbyallusersinthedatabase.The
sameistrueforPUBLICSYNONYMS.
Inourearlierexample,theuserSYSTEMcouldhavegivenuserLISAthe
systemprivilegetocreatepublicsynonymsbyissuingthefollowingstatement:
Then,later,theuserLISAcouldhaveusedthatsystemprivilegetocreatea
PUBLICSYNONYMlikethis:
Finally,onceuserHENRYgotaroundtoissuingDMLstatementsonthe
WEBINARStable,HENRYcouldhaveomittedtheschemaprefixandinstead
simplyexecutedthisstatement:
Inthisinstance,HENRYwouldbespecifyingtheWEBINARSobject
PUBLICSYNONYM,whichinturnpointstotheobjectLISA.WEBINARS.
NotethatnoobjectprivilegehadtobegrantedonthePUBLICSYNONYM
objecttoHENRY.AllobjectsownedbyPUBLICareautomaticallyavailable
andaccessibletoallusersinthedatabase,presentandfuture.However,
privilegesmustbegrantedtowhateverobjectforwhichthePUBLIC
SYNONYMservesasanalias.It’sonethingtohaveprivilegesonaPUBLIC
SYNONYMthatreferencesatable,butit’sanotherthingtohaveprivilegeson
thetableitreferences.Allusershaveprivilegesautomaticallyonanyobject
ownedbyPUBLIC;theydonothaveautomaticallygrantedprivilegeson
anythingaPUBLICSYNONYMreferences;suchprivilegesmustbegranted
explicitly.
ThissortofusageisthemostcommonpurposeofthePUBLICSYNONYM
object.
NotethattocreatePUBLICSYNONYMobjects,auseraccountmusthave
theCREATEPUBLICSYNONYMsystemprivilege.
NamePriority,Revisited
YoumayrecallourdiscussioninChapter2aboutaconceptcallednamespace.
Whenausermakesareferencetoanobjectbyname,SQLwillusethatnameto
searchforthatobjectasfollows:
First,SQLlooksinthelocalnamespace,whichcontainsobjects
ownedbytheuseraccount:tables,views,sequences,privatesynonyms,
andsomethingcalleduser-definedtypes(theseobjectsarebeyondthe
scopeoftheexam).
Next,SQLlooksinthedatabasenamespace,whichcontainsusers,
roles,andpublicsynonyms.
ThisconceptwasdemonstratedgraphicallyinFigure2-2.
WITHGRANTOPTION
Ifyouwanttograntanotheruseraparticularobjectprivilegeandincludethe
abilityfortheusertograntthatsameobjectprivilegetoyetanotheruser,then
includetheWITHGRANTOPTIONclauseintheGRANTstatement.Here’san
example:
ThisgrantgivesuserHENRYtheabilitytoissueSELECTandUPDATE
statementsontableWEBINARS,alongwiththeabilitytograntthoseprivileges
tootherusers.HENRYisnotobligatedtograntthesetofprivilegestogether;
HENRYcanchoosetobeselective.
NowuserHAROLDhastheabilitytoissueSELECTstatementson
LISA.WEBINARS,aswellastheabilitytograntthatprivilegetoothers.But
HENRYdidnotpassalongtheUPDATEprivilege.
REVOKE
UserLISAmaychoosetorevokeprivilegesfromHENRY,likethis:
IfuserLISAdoesthis,thenHENRYandHAROLDloseallprivileges,as
doesanyonetowhomtheyextendedprivilegeswiththeirWITHGRANT
OPTIONoption.
Inotherwords,therevocationofobjectprivileges“cascades.”Notethatthis
isdifferentfromsystemprivilegerevocation,whichdoesnotcascade,aswe
statedearlierinthischapter.
NotethattheREVOKEstatementdoesnotrequiretheWITHGRANT
OPTIONclause.REVOKEdoesn’tcarewhetherthatoptionhadbeenincluded;
itjustrevokesallspecifiedprivilegesandcascadesthechangethroughoutall
useraccountsasrequired.
ALLPRIVILEGES
TheALLPRIVILEGESoptionworkswithgrantingandrevokingobject
privilegesinmuchthesamewayitdoeswithsystemprivileges,withsome
differences.Here’sanexample:
ThisstatementgivesallprivilegesontheobjectWEBINARStoHENRY,
exceptfortheabilitytograntprivileges.Togranttheabilitytogrant,usethis:
ThekeywordPRIVILEGESisnotrequiredwhengrantingobjectprivileges.
ThesameistruewithREVOKEwhenusedwithobjectprivileges.
Thefollowingisalsogood:
Thisshorthandwayofrevokingobjectprivilegessparestheeffortof
identifyingalltheindividualobjectprivilegesthatmayhavealreadybeen
grantedtoHENRYontheWEBINARStableandrevokesthemallatonce.
NotethatthekeywordPRIVILEGESisoptionalwhenworkingwithobject
privileges,butnotwhenworkingwithsystemprivileges.
IfyouuseREVOKEALLtorevokeobjectprivilegesfromauserandno
objectprivilegesexistontheobjectforthatuser,thennoerrormessageresults,
andthestatementexecutessuccessfullywithnopracticaleffect.
DependentPrivileges
IfuserAownsaview,whichisbasedonatablethatuserAalsoowns,anduser
AgrantsprivilegesontheviewtouserB,thenuserBcanaccesstheview
withoutprivilegestotheunderlyingtable.
IfuserAcreatesatableandapublicsynonym,thenuserBhasimmediate
visibilityofthepublicsynonymbecausethesynonymisownedbyPUBLICand
allusershavevisibilityofallobjectsownedbyPUBLIC.However,userBstill
requiresprivilegesonthetableforwhichthepublicsynonymisanalias.Ifthe
publicsynonymreferencesaviewthatuserAowns,thenuserBmusthave
objectprivilegesontheview,butisnotrequiredtohaveaccesstoitsunderlying
table.
Ifyougrantprivilegesonatableandthendropthetable,the
privilegesaredroppedwiththetable.Ifyoulaterre-createthetable,you
mustalsogranttheprivilegesagain.However,ifyourestoreadropped
tablewiththeFLASHBACKTABLE…BEFOREDROPstatement,you
willrecoverthetable,itsassociatedindices,andthetable’sgranted
privileges,andyouwillnotneedtogranttheprivilegesagain.
ViewPrivilegesintheDataDictionary
We’vealreadylookedatthedatadictionaryandseenhowitprovides
informationaboutthestateofobjectsinthedatabase,aswellasprovidingsome
historicinformation.
Therearemanyviewsinthedatadictionarythatpresentinformationabout
systemprivilegesandobjectprivileges.SeeTable14-3foralistingofsomeof
theseviews.
TABLE14-3 DataAboutPrivilegesintheDataDictionary
Forexample,toseewhatsystemprivilegesaregrantedtoyourcurrentuser
account,youcanquerythedatadictionaryviewUSER_SYS_PRIVS.Here’s
whattheresultsmightlooklikefromuseraccountLISA:
TheequivalentdatadictionaryviewDBA_SYS_PRIVSallowsyoutoseethe
sameinformationforotherusers.
Aprivilegemaybegranteddirectlyasaprivilegeorindirectlyaspart
ofarole.Ifyouintendtodropaprivilegefromauser,usethedata
dictionarytodeterminewhetherthatsameprivilegeisgrantedtoarole
thatisalsograntedtothesameuser.Ifso,thentheprivilegeyoudropped
directlyisstillgrantedtotheuserindirectlythroughtherole.
Toseealltheobjectprivilegesthatyourcurrentuseraccountmayhave
grantedtoothersorthatmayhavebeengrantedbyothers,youcanusethe
followingquery.Thisiswhattheresultsmightlooklikewithintheuseraccount
LISA:
NotethatthefirstseveralrowsshowobjectprivilegesgrantedbyEFCODD
toLISA.ThefinaltworowsshowobjectprivilegesgrantedbyLISAtoHENRY.
Thesearejustafewexamplesofthesortofinformationthedatadictionary
providesaboutsystemprivilegesandobjectprivilegesthathavebeengrantedto
andfromuseraccountswithinthedatabase.
WheninspectingdatadictionaryviewslikeDBA_TAB_PRIVSor
DBA_SYS_PRIVStoseewhatprivilegeshavebeengrantedtoaparticular
useraccount,youcanchecktheGRANTEEcolumnfortheappropriateUSER
name.However,don’tforgettoalsocheckforrowswhereGRANTEE=
'PUBLIC';theseprivilegesarealsoavailabletoyouruseraccount.
GrantRoles
Aroleisadatabaseobjectthatyoucancreateandtowhichyoucanassign
systemprivilegesand/orobjectprivileges.Youcanalsoassignotherrolestoa
givenrole.Onceitiscreated,youcangrantaroletoauserjustasyoucangrant
privilegestoauser.Theuseristhenautomaticallygrantedanyprivileges
containedwithintherole.Aroleisanexcellentwaytomanagethevarious
privilegesrequiredforperformingdifferenttasksinthedatabaseandtoorganize
theprocessofgrantingandrevokingprivileges.
Youmaygranttheroletoasmanyuseraccountsasyouwant.
Ifanyprivilegeissubsequentlyrevokedfromtherole,itisalsorevokedfrom
anyuserstowhomtherolehasbeengranted.Inotherwords,changestoroles
cascadetotheuserstowhomtheroleisgranted.
Threerolesinparticularhavehistoricallybeenassociatedwithstandard
Oracledatabases,buttheyarebeingphasedout.Onapracticallevel,though,it’s
goodtoknowaboutthem,ifyoudon’talready.ThethreerolesareCONNECT,
RESOURCE,andDBA.TheCONNECTroleconsistsoftheCREATE
SESSIONsystemprivilege,intendedforthetypicalgenericenduser.
RESOURCEisacollectionofsystemprivilegesintendedforthetypical
applicationdeveloper.DBAisintendedforthetypicaldatabaseadministrator.
EachcanbeseenindetailinthedatadictionaryviewDBA_SYS_PRIVS(see
Table14-4fordetails).Allthreerolesarestillincludedineachimplementation
oftheOracledatabaseasofthiswriting,butOraclehasstatedformallythatthe
useoftheserolesisnowofficiallydiscouraged,andtheirinclusioninfuture
databaseimplementationsisnotguaranteed.OracleCorporation’sofficial
positionisthatyoushouldcreateyourownsetofrolesasrequired.
YoucanrefertothedatainTable14-4togetanideaofthekindofsystem
privilegesyoumaywanttoincludeinyourROLEobjects.
TABLE14-4 TheClassicRolesCONNECT,RESOURCE,andDBA
Tocreatearole,auseraccountneedstheCREATEROLEsystemprivilege.
Forexample,theuseraccountEFCODDownsseveraltablesandwantsto
grantprivilegesonthesetablestosomeusersinthedatabase.Someofthese
userswillbeperformingqueriesonthetablesandnothingmore.Otherswillbe
responsibleforperformingchangestothedata.Therefore,wewanttocreatetwo
differentrolesandgrantthemthenecessaryprivileges.
Intheprecedingcode,wecreatetwoROLEobjects:onecalled
CRUISE_ANALYST,towhichwegrantsomeSELECTprivilegesontables,and
anothercalledCRUISE_OPERATOR,towhichwegrantsomeotherprivileges.
Oncetheyarecreated,wecangranttheserolestouseraccountsinthedatabase.
Oncearoleisgranted,auserhasaccesstoalloftheprivilegeswithinit.
Arolecanbegrantedtoanotherrole.
ArolecanbegrantedWITHADMINOPTIONtoempowertherecipientto
granttheroletoyetanotheruser.Here’sanexample:
IfausergrantsaroletoanotheruserandusesWITHADMINOPTION,the
secondusermayfurthergrantthesameroletoathirduser.Ifthefirstuser
revokestherolefromtheseconduser,thethirduserretainstheroleuntilitis
explicitlyrevokedfromthethirduserbyaqualifieduser.
Table14-5listssomeofthedatadictionaryviewsthatprovideinformation
aboutexistingrolesinthedatabase.
TABLE14-5 DataDictionaryViewswithInformationAboutROLEObjects
Rolesexistinanamespacethatresidesoutsideofanyuseraccount.
Therefore,youcancreateroleswithnamesthatarethesameasobjectswithina
useraccount,suchastablesandviews.That’snotnecessarilyagoodidea,but
it’sallowedinthedatabase.
Auseraccountmaybegrantedmultiplerolesatonce.
Let’ssayyoucreateanobject,thengrantaprivilegeonthatobjecttoa
role,andthengranttheroletoauser.Ifyoudroptheobject,thenyou
alsodropthegrantedobjectprivilegetotherole.However,therolestill
exists,andthegrantoftheroletotheuserstillexists.Ifyousubsequently
re-createtheobjectandthengranttheobjectprivilegetotheroleonce
again,thenyou’vere-createdthesituationbeforetheobjectwasdropped.
Inotherwords,youdonotneedtore-createtheroleorgranttheroleto
theuseronceagainsinceneitherwasaffectedbytheactofdroppingthe
objectonwhichtheprivilegehadoriginallybeengranted.
CERTIFICATIONOBJECTIVE14.03
DistinguishBetweenPrivilegesandRoles
Aroleobjectdoesnotrepresentprivilegesinandofitself.Itismerelya
collectionofzeroormoreprivileges.Aroleexistsindependentlyofthe
privilegesitmay—ormaynot—contain.Furthermore,therelationshipauser
accounthastoagrantedroleisseparatefromanyprivilegesthatmayhavebeen
granteddirectlytotheuseraccount.Inotherwords,ifauseraccountalreadyhas
anyobjectprivilegesgranteddirectlytoitasaresultofearlierGRANT
statementsandthenlaterisgrantedarolethatduplicatesanyofthoseprivileges,
thentheroleexistsseparatelyfromthoseoriginallygrantedprivileges,which
existindependentlyoftherole.Iftheroleislaterrevoked,thatrevocationdoes
notadverselyaffectanyseparatelygrantedprivilegesgivendirectlytotheuser
account.
IfuserHENRYwerealreadygrantedaprivilegethathappenstobeduplicated
withintheroleCRUISE_ANALYSTandthensubsequentlytheroleisgranted
butthenisrevoked,likethis:
thenanyobjectprivilegesgranteddirectlytoHENRYstillexist.
Forexample,examinethefollowingcode:
UserHENRYstillhasSELECTonINVOICESbecauseofline1,inspiteof
lines2through5.
Similarly,iftheroleisrestoredbutthedirectobjectprivilegeisrevoked,
HENRYstillhasaccessthroughtherole.Here’sanexample:
Rememberthat“privileges”mayrefertoeithersystemprivilegesor
objectprivileges,whichareverydifferent.
Rolesconsistofsomecombinationofoneormoresystemand/orobject
privilegesand/orotherroles.
HENRYstillhasprivilegesonINVOICESinspiteofline5.Thereasonisthe
CRUISE_ACCOUNTANTrole,fromlines2through4.
However,iftheobjectprivilegerevokedfromHENRYinline5werealsoto
berevokedfromtheCRUISE_ACCOUNTANTrole,thentheobjectprivilege
wouldberemovedfromHENRYaltogether.
CERTIFICATIONSUMMARY
Asystemprivilegeistherighttoperformataskinthedatabase,usingaDDL,
DCL,orDMLstatementonobjectsingeneral.Therighttoperformthosetasks
onaparticularobjectinthedatabaseisanobjectprivilege.Finally,arole
combinesprivilegesintoasingleobjectsothatacombinationofprivilegescan
bemanagedasagroup.
TheSQLstatementsGRANTandREVOKEareusedtoissuesystem
privilegesandobjectprivilegesandtotakethemaway.Privilegesaregivento—
ortakenawayfrom—useraccounts.Anyuserinthedatabasemusthave
privilegestoperformanytask.TheactoflogginginrequirestheCREATE
SESSIONsystemprivilege.OtherprivilegesincludeCREATEPUBLIC
SYNONYMandCREATETABLE.
TheANYkeywordinasystemprivilegeindicatestheabilitytoworkwith
objectsthatareownedbyanyuseraccount.
Auseraccount,bydefault,hasobjectprivilegesontheobjectsitowns.
Objectprivilegesarerequiredforausertobeabletointeractwithobjectsitdoes
notown.
Insteadofgrantingprivilegestoauser,youmaycreatearole,thengrant
privilegestoarole,andthengranttheroletooneormoreusers.Theadvantage
isthatifyouhavemultipleusers,aroleismucheasiertochangesinceyoucan
grantorrevokeprivilegesasdesiredaftertherolehasbeenassignedtoany
numberofusers,andalloftheuserswillautomaticallyhavethenewprivileges
grantedorrevokedautomatically.
Thedatadictionaryprovidesinformationaboutsystemprivileges,object
privileges,androlesfromtheperspectiveofboththegrantorandthegrantee.
✓TWO-MINUTEDRILL
DifferentiateSystemPrivilegesfromObjectPrivileges
TherighttouseanygivenSQLstatementand/ortogenerally
performataskinthedatabaseisasystemprivilege.
Therighttoperformsometaskonaspecificexistingobjectinthe
databaseisanobjectprivilege.
Bothsystemandobjectprivilegesaregrantedtoandrevokedfrom
usersinthedatabase.
SystemprivilegesmaybegrantedWITHADMINOPTION,which
providestheabilityfortherecipienttograntthesameprivilegetoyet
anotheruser.
Whenasystemprivilegeisrevoked,therevocationdoesnot
cascade,meaningthatitisrevokedonlyfromtheuserfromwhomitis
beingrevoked,notfromotheruserstowhomtherevokedusermayhave
extendedtheprivilege.
TheALLPRIVILEGESkeywordscanbeusedtograntorrevoke
allprivilegestoorfromauser.
GrantPrivilegesonTablesandonaUser
ObjectprivilegescorrespondtoDMLstatementsandtoDDL
statementsthatarerelevanttoexistingobjects.
ObjectprivilegesmaybegrantedWITHGRANTOPTION,which
providestheabilityfortherecipienttograntthesameprivilegetoyet
anotheruser.
Whenanobjectprivilegeisrevoked,therevocationcascades,
meaningthatitisrevokedfromtheuserfromwhomitisbeingrevoked,
aswellasfromotheruserstowhomtherevokedusermayhaveextended
theprivilege.
Whenauserhasbeengrantedaccesstoanobject,theobjectname
willrequireaschemanameprefixtobecorrectlyidentified.
APUBLICSYNONYMcanprovideanalternativenameforthe
schema-prefixedversionofthegrantedobject.
TheALLPRIVILEGESkeywordscanbeusedtograntorrevoke
allprivilegestoorfromauser.
Avarietyofdatadictionaryviewsprovideinformationabout
systemandobjectprivileges.
Usersmayseeprivilegesgrantedtothem,orgrantedbythemto
others,byqueryingthedatadictionary.
AroleiscreatedwiththeCREATEROLEstatement.
RolesmaybegrantedWITHADMINOPTION,whichprovidesthe
abilityfortherecipienttograntthesameroletoyetanotheruser.
Rolesexistinanamespaceoutsideofanindividualuseraccount.
Aroleisacollectionofprivilegesandotherroles.
Arolemaybegrantedtoanotherrole.
DistinguishBetweenPrivilegesandRoles
Aprivilegegranteddirectlytoauserexistsindependentlyfroma
privilegegrantedtoarole.
Ifyourevokeaprivilegedirectlyfromauserwhoalsohasbeen
grantedarolecontainingthesameprivilege,theroleremains
unchanged,andtheuserstillhasprivilegesbywayoftherole.
Thesamesituationistruewithregardtorevokingprivileges
directlyfromroles;ifyourevokeaprivilegefromarolethatauser
alreadyhasthroughadirectgrant,thedirectgrantstaysinforce.
SELFTEST
Thefollowingquestionswillhelpyoumeasureyourunderstandingofthe
materialpresentedinthischapter.Chooseoneanswerforeachquestion,unless
otherwisedirected.
DifferentiateSystemPrivilegesfromObjectPrivileges
1.WhichofthefollowingSQLstatementswillauthorizetheuseraccount
JESSEtocreatetablesineachandeveryuseraccountinthedatabase?
A.GRANTCREATEALLTABLETOJESSE;
B.GRANTCREATEPUBLICTABLETOJESSE;
C.GRANTCREATEANYTABLETOJESSE;
D.GRANTCREATETABLETOJESSEWITHPUBLIC
OPTION;
2.YouareloggedintouseraccountFREDandhavebeentaskedwith
grantingprivilegestotheuseraccountETHEL.Youexecutethefollowing
SQLstatements:
Assumingbothstatementsexecutesuccessfully,whatistheresult?
A.ETHELdoesnothavethesystemprivilegeCREATEANY
TABLEortherighttogranttheCREATEANYTABLEsystem
privilegetoanyotheruser.
B.ETHELhasthesystemprivilegeCREATEANYTABLE
becausetheWITHADMINOPTIONclausewasn’tincludedinthe
REVOKEstatement.
C.ETHELnolongerhasthesystemprivilegeCREATEANY
TABLEbutstillhastherighttogranttheCREATEANYTABLE
systemprivilegetoanyotheruser,sincetheWITHADMINOPTION
clausewasomittedfromtheREVOKEstatement.However,ETHEL
maynotgranttheCREATEANYTABLEprivilegetoherself.
D.ETHELnolongerhasthesystemprivilegeCREATEANY
TABLEbutstillhastherighttogranttheCREATEANYTABLE
systemprivilegetoanyotherusersincetheWITHADMINOPTION
clausewasomitted.Furthermore,ETHELmaygranttheCREATE
ANYTABLEprivilegetoherselfbecauseoftheWITHADMIN
OPTIONclause.
3.Whichofthefollowingisthesystemprivilegethatisrequiredasa
minimumtoallowauseraccounttologintothedatabase?
A.CREATEANYLOGIN
B.CREATEANYSESSION
C.CREATESESSION
D.CREATETABLE
4.Whichofthefollowingisthesystemprivilegethatempowersthe
granteetocreateanindexinhisorherownuseraccountbutnotinthe
accountsofothers?
A.CREATETABLE
B.CREATEANYTABLE
C.CREATEINDEX
D.CREATEANYINDEX
GrantPrivilegesonTablesandonaUser
5.YouruseraccountownsatableBACK_ORDERS,andyouwantto
grantprivilegesonthetabletoauseraccountnamedCARUSO,which
alreadyhasthesystemprivilegesCREATESESSIONandUNLIMITED
TABLESPACE.ExaminethefollowingSQLstatement:
Oncethisstatementhasbeenexecuted,whichofthefollowing
statementswillbetrueforuserCARUSO?
A.CARUSOwillhaveSELECTprivilegesonBACK_ORDERS
butnottheabilitytogiveotherusersSELECTprivilegeson
BACK_ORDERS.
B.CARUSOwillhaveSELECTprivilegesonBACK_ORDERS,as
wellastheabilitytogiveotherusersSELECTprivilegeson
BACK_ORDERS.
C.CARUSOwillhaveSELECT,INSERT,UPDATE,andDELETE
privilegesonBACK_
ORDERSbutnottheabilitytogiveotherusersthosesameprivileges
onBACK_ORDERS.
D.CARUSOwillhaveSELECTandALTERTABLEprivilegeson
BACK_ORDERSbutnottheabilitytogiveotherusersthosesame
privilegesonBACK_ORDERS.
6.Youruseraccountownsanupdatableview,BACKLOG,whichis
basedonthetablePROJECTS.YouaretaskedtogiveSELECTand
UPDATEcapabilitiestoanotheruseraccountnamedMARINO.Currently,
MARINOhasnoprivilegesoneitherthetableortheview.Youwantfor
MARINOtohavetheabilitytograntSELECTontheviewtootherusersas
well.ExaminethefollowingSQLcode:
Whichofthefollowingstatementsistrue?
A.Thestatementswillfail,andMARINOwillnotbeabletouse
theview.
B.Thestatementswillexecutesuccessfully,butMARINOwillnot
beabletoSELECTfromtheviewbecausethePROJECTStablehas
notbeengrantedtoMARINO.
C.Thestatementswillexecutesuccessfully,andMARINOwillbe
abletoSELECTfromtheviewbutnotUPDATEtheview.
D.Thestatementswillexecutesuccessfullyandperformas
intended.
7.UseraccountMUSKIEownsatablecalledCBAY.Whichofthe
followingstatementscanbeexecutedbyMUSKIEandenableuser
ONEILLtoexecuteUPDATEstatementsontheCBAYtable?(Choose
three.)
A.GRANTALLONCBAYTOONEILL;
B.GRANTALLPRIVILEGESTOONEILL;
C.GRANTALLTOONEILL;
D.GRANTINSERT,UPDATEONCBAYTOONEILL;
8.Examinethefollowingtwoclaims:
[1]TheDBA_TAB_PRIVSdatadictionaryviewallowsauser
accounttoseeobjectprivilegesithasgrantedtootheruseraccounts.
[2]TheDBA_TAB_PRIVSdatadictionaryviewallowsauser
accounttoseeobjectprivilegesgrantedbyotheruseraccountsto
itself.
Whichoftheseclaimsistrue?
A.Only1
B.Only2
C.Both1and2
D.Neither1nor2
9.Whichofthefollowingdatadictionaryviewscontainsinformation
aboutgrantsontablesthathavebeenmadebyotheruserstoyouruser
account,aswellasgrantsontablesthathavebeenmadebyyouruser
accounttootheruseraccounts?
A.USER_TAB_COLUMNS
B.USER_TAB_PRIVS
C.USER_TABLES
D.ALL_TAB_PRIVS_RECD
10.Whatcanbegrantedtoarole?(Chooseallthatapply.)
A.Systemprivileges
B.Objectprivileges
C.Roles
D.Noneoftheabove
11.WhichofthefollowingstatementswillgranttheroleOMBUDSMANtouser
JOSHUAinsuchawaythatJOSHUAmaygranttheroletoanotheruser?
A.GRANTOMBUDSMANTOJOSHUAWITHADMIN
OPTION;
B.GRANTOMBUDSMANTOJOSHUAWITHGRANT
OPTION;
C.GRANTOMBUDSMANTOJOSHUAWITHROLEOPTION;
D.GRANTOMBUDSMANTOJOSHUACASCADE;
12.UserHARDINGownsatableTEAPOT.UserHARDINGthenexecutesthe
followingSQLstatementstogiveaccesstothetabletouserALBERT:
WhichofthefollowingstatementscanuserALBERTnowexecuteon
theTEAPOTtable?
A.SELECT*FROMDOME.HARDING.TEAPOT;
B.SELECT*FROMHARDING.DOME.TEAPOT;
C.SELECT*FROMHARDING.TEAPOT;
D.Noneoftheabove
DistinguishBetweenPrivilegesandRoles
13.Arole:
A.Takestheplaceofprivilegesautomaticallysothatanyprivilege
grantedtoarolesupersedesanygrantsthathavealreadybeengranted
directlytoauser
B.Cannotbegiventhesamenameasatable
C.Canbegrantedtoauser,whocanbegrantedonlyoneroleata
time
D.CanbecreatedbyauseronlyifthatuserhastheCREATE
ROLEsystemprivilege
14.YouhaveatableFURNISHINGSandaretoldtograntDELETEprivileges
onthetabletouserHEARST.ExaminethefollowingSQLstatements:
NowyouaretoldtochangetheprivilegesgiventoHEARSTsothat
HEARSTcannolongerexecuteDELETEstatementsonthe
FURNISHINGStable.Whichofthefollowingwillaccomplishthegoal?
(Choosethebestanswer.)
A.REVOKEDELETEONFURNISHINGSFROMHEARST;
B.REVOKEDELETEONFURNISHINGSFROMMGR;
C.REVOKEDELETEONFURNISHINGSFROMHEARST,
MGR;
D.Noneoftheabove
15.Assumeadatabasewiththreevalidusers:NEIL,BUZZ,andMICHAEL.
Assumeallusershavetheappropriateprivilegestheyrequiretoperform
thetasksshownhere.AssumeNEILownsatablecalledPROVISIONS.
Examinethefollowingcode(assumeallpasswordreferencesarevalid):
Whatobjectisidentifiedinline11bythenamePROVISIONS?
A.Thepublicsynonymcreatedinline7
B.Thesynonymcreatedinline10
C.Nothing,becauseuserNEILdidnotincludeWITHGRANT
OPTIONSintheGRANTSELECTONPROVISIONSTOBUZZ
statement
D.Somethingelsenotlistedabove
SELFTESTANSWERS
DifferentiateSystemPrivilegesfromObjectPrivileges
1. C.ThesystemprivilegeCREATEANYTABLEisthesystem
privilegethatyou’relookingforinthisquestion.ThekeywordANYis
foundinmanysystemprivilegestoindicatethattheuserauthorizedwith
thesystemprivilegemayperformthetaskasthoughitwereanyuser
accountinthedatabase.
A,B,andDareincorrect.ThereisnoALLkeywordinthiscontext,and
PUBLICdoesnotapplyhere.ThereisnosystemprivilegewiththeWITH
PUBLICOPTIONkeywords.
2. A.TheWITHADMINOPTIONclauseisnotallowednorneededin
theREVOKEstatement.
B,C,andDareincorrect.Theyareallinterestingideas,buttheyareall
wrong.
3. C.TheCREATESESSIONsystemprivilegeistheminimum
requirement.
A,B,andDareincorrect.ThereisnosystemprivilegeCREATEANY
LOGINorCREATEANYSESSION.CREATETABLEisnotrequiredto
establishausersession.
4. A.TheCREATETABLEprivilegealsoincludestheabilitytocreate
anindex.RememberthataCREATETABLEstatementmayincludethe
PRIMARYKEYorUNIQUEconstraint,which—ifcreated—will
automaticallycausethecreationofanindextosupporteachconstraint.
B,C,andDareincorrect.Thereisn’taCREATEINDEXsystem
privilege.TheabilityisincludedwithCREATETABLE.CREATEANY
TABLEempowersthegranteetheabilitytocreatetablesintheaccountsof
others,whichpotentiallymayalsocreateindicesinthosesameaccounts.
CREATEANYINDEXisavalidsystemprivilegeforcreatingindex
objectsinuseraccountsotherthanyourown.
GrantPrivilegesonTablesandonaUser
5. A.GRANTSELECTONtableTOusergivestheusertheabilityto
SELECTonthetableandnothingmore.
B,C,andDareincorrect.TogiveCARUSOtheabilitytoSELECTon
thetableaswellastograntotherusersSELECT,theWITHGRANT
OPTIONclausewouldneedtohavebeenincludedwiththeGRANT
statement,asinGRANTSELECTONBACK_ORDERSTOCARUSO
WITHGRANTOPTION.TogranttheotherDMLstatementsonthetable,
eachwouldhavetohavebeenincluded,asinGRANTSELECT,INSERT,
UPDATE,DELETEONBACK_ORDERSTOCARUSO.Togrant
SELECTandALTER,bothwouldhavetohavebeennamed,asinGRANT
SELECT,ALTERONBACK_ORDERSTOCARUSO.
6. D.Thestatementsaresyntacticallycorrectandwillperformas
intended.
A,B,andCareincorrect.ThePROJECTStabledoesnotneedtobe
grantedtoMARINOsincetheVIEWhasbeengranted.SincetheVIEWis
updatable,thentheUPDATEprivilegewillworkaswell.
7. A,B,andD.AllthreeformsresultintheUPDATEprivilegebeing
grantedtouserONEILLfortheCBAYtable.
Cisincorrect.ThisstatementisaninvalidSQLstatement.Iteitherneeds
forthekeywordPRIVILEGEStograntallsystemprivilegestoONEILLor
needstonameanobjectforwhichALLprivilegesshouldbegranted.The
questionisspecificallyaskingaboutgrantingprivilegesontheCBAYtable,
sotheALLPRIVILEGESformwouldnotwork.
8. C.ThedatadictionaryviewDBA_TAB_PRIVSallowsausertosee
privilegesthathavebeengrantedtoitselforbyitselftoothers.
A,B,andDareincorrect.
9. B.USER_TAB_PRIVSisthecorrectanswer.
A,C,andDareincorrect.USER_TAB_COLUMNShasnoinformation
aboutgrants.NeitherdoesUSER_TABLES.The
ALL_TAB_PRIVS_RECDviewcontainsdataaboutincominggrantsonly.
10. A,B,andC.Bothsystemandobjectprivileges,aswellasotherroles,can
begrantedtoanygivenrole.
Disincorrect.
11. A.WITHADMINOPTIONiswhatisusedforroles.
B,C,andDareincorrect.WITHGRANTOPTIONworksforobject
privilegesbutnotroles.ThereisnosuchclauseasWITHROLEOPTION.
CASCADEdoesnotapplyhere.
12. C.Theschemanameprefixcorrectlyidentifiesthetable.Inaddition,since
thepublicsynonymTEAPOTreferencesthetable,thenDESCTEAPOT
wouldalsohaveworked—butthatwasnotoneoftheoptionslisted.
A,B,andDareincorrect.Youcannotusetheroleasaprefixorany
othercomponentofthenameofadatabaseobject.
DistinguishBetweenPrivilegesandRoles
13. D.TheCREATEROLEprivilegeisrequiredtocreatearole.
A,B,andCareincorrect.Aroledoesnotreplaceprivilegesbutinstead
isgrantedalongsideofthem.Arolemaybeusedtoreplaceprivilegesasa
managementchoice,andinfactsuchanapproachisadvisable,butitisnot
doneautomatically.Rolesexistinadifferentnamespacefromtablesand
mayduplicatetablenames.Ausermaybegrantedmultiplerolesatany
giventime.
14. C.ThisSQLstatementaccomplishesthegoalinonestatement.
A,B,andDareincorrect.AandBarehelpfulbutdonotcompletely
accomplishthetask.DisincorrectbecauseCiscorrect.
15. B.FromwithintheMICHAELuseraccount,SQLfirstsearchesthelocal
namespaceandthensearchesthedatabasenamespace.Thelocal
namespacecontainstheprivatesynonym,andthatwillbefoundfirst,
beforeSQLlooksinthedatabasenamespace.
A,C,andDareincorrect.TheGRANTstatementissuedbyNEILdoes
notrequireWITHGRANTOPTIONforthesynonymstofunction.
T
A
AbouttheDownload
hise-bookcomeswithTotalTestercustomizablepracticeexamsoftwarewith
140practiceexamquestions.TheTotalTestersoftwarecanbedownloaded
andinstalledonanyWindowsVista/7/8/10computerandmustbeinstalledto
accesstheTotalTesterpracticeexams.
TodownloadtheTotalTester,simplyclickthelinkbelowandfollowthe
directionsforfreeonlineregistration.
http://www.totalsem.com/1259584615d
SystemRequirements
ThesoftwarerequiresWindowsVistaorhigherand30MBofharddiskspacefor
fullinstallation,inadditiontoacurrentorpriormajorreleaseofChrome,
Firefox,InternetExplorer,orSafari.Torun,thescreenresolutionmustbesetto
1024×768orhigher.
InstallingandRunningTotalTester
Onceyou’vedownloadedtheTotalTestersoftware,double-clicktheSetup.exe
icon.Thiswillbegintheinstallationprocessandplaceanicononyourdesktop
andinyourStartmenu.TorunTotalTester,navigatetoStart|(All)Programs|
TotalSeminarsordouble-clicktheicononyourdesktop.
TouninstalltheTotalTestersoftware,gotoStart|ControlPanel|Programs
AndFeatures,andthenselecttheTotalTesterprogram.SelectRemove,and
Windowswillcompletelyuninstallthesoftware.
AboutTotalTester
TotalTesterprovidesyouwithasimulationoftheOracleDatabaseSQLexam
(Exam1Z0-071).ExamscanbetakeninPracticeMode,ExamMode,orCustom
Mode.PracticeModeprovidesanassistancewindowwithhints,referencesto
thebook,explanationsofthecorrectandincorrectanswers,andtheoptionto
checkyouranswerasyoutakethetest.ExamModeprovidesasimulationofthe
actualexam.Thenumberofquestions,thetypesofquestions,andthetime
allowedareintendedtobeanaccuraterepresentationoftheexamenvironment.
CustomModeallowsyoutocreatecustomexamsfromselecteddomainsor
chapters,andyoucanfurthercustomizethenumberofquestionsandtime
allowed.
Totakeatest,launchtheprogramandselectOracleDatabaseSQLfromthe
InstalledQuestionPackslist.YoucanthenselectPracticeMode,ExamMode,or
CustomMode.Allexamsprovideanoverallgradeandagradebrokendownby
domain.
TechnicalSupport
ForquestionsregardingtheTotalTestersoftwaredownloadoroperation,visit
www.totalsem.comore-mailsupport@totalsem.com.
Forquestionsregardingthee-bookcontent,pleasee-mailhep_customer-
service@mheducation.com.ForcustomersoutsidetheUnitedStates,e-mail
international_cs@mheducation.com.
Glossary
1GLFirst-generationlanguage.The1sand0sthatcomputersuseto
communicate.Binarylanguage.
2GLSecond-generationlanguage.Assemblerlanguage.
3GLThird-generationlanguage.Ageneralcategoryofcomputerprogramming
languagesthattendtosupportstructuredorobject-orientedprogrammingina
mannerthatisclosertothespokenwordthan2GLs.Common3GLsareJava,
C++,JavaScript,andPHP.
4GLFourth-generationlanguage.Closertothespokenwordthan3GLs.The
mostwell-knownandwidelyused4GLisSQL.
administratorSeedatabaseadministrator.
aggregateAsinglevaluerepresentinganynumberofothervalues.
aliasAnalternativenameforsomething.Forexample,JoeisanaliasforJoseph.
alphabeticDescribesthelettersofthealphabet.
alphanumericDescribesthelettersofthealphabetandnumbers.
ALTERASQLstatementthatmodifiesthestructure,thename,orsomeother
attributeofanexistingobjectinthedatabase.(Note:Thereareexceptionstothis
definitionthatoccurwhenALTERiscombinedwiththekeywordSESSIONor
STATEMENT.)analyticfunctionsFunctionsthatoperateonagroupofrows
(knownasawindow)andreturnmultiplerowsforeachgroup.Sometimescalled
analyticalfunctions.
ANSIAmericanNationalStandardsInstitute.Anorganizationthatoverseesa
numberofvoluntarycommitteesthatsetstandardsformanyindustries,
includingsoftwaredevelopmentandinformationtechnology.
attributeApropertyorcharacteristic.Examplesmightincludeaname,ZIP
code,orentrydate.Correspondstoacolumninatable.Seealsoentity.
BLOBBinaryLargeObject.Adatatypethatstoresunstructuredbinarydata,up
to128terabytes.BLOBdatatypescanberolledbackorcommittedaspartof
anydatabasetransaction.Suitableforstoringmultimediadata.
BooleanReferstothevaluationofexpressionsaseithertrue,false,orunknown
andusingthelogicaloperatorsAND,OR,andNOT.Namedafterthe
mathematicianGeorgeBoole.
built-inAlreadypresent.SQLbuilt-infunctionsarethosethatcomealready
installedinadatabase,asopposedtouser-definedfunctionsthatyoucancreate
yourselfandaddtothesetofavailablefunctionsinadatabase.
CartesianproductThecombinationofeachrowinonetablewitheveryrowin
anothertable.Theresultoftwoormoretablesjoinedtogetherwithnospecified
joincriteria.Alsoknownasacross-join.
caseinsensitiveWithoutregardforwhetheraletterisinuppercaseorlowercase
form.Forexample,whenperformingacase-insensitivecomparisonoftheletter
Aandthelettera,thetwoareequal.
casesensitiveWithregardforwhetheraletterisinuppercaseorlowercaseform.
Forexample,whenperformingacase-sensitivecomparisonoftheletterAand
thelettera,thetwoarenotequal.
characterThesymbolsofawritingsystem.
characterclassAlsoknownasPOSIXcharacterclasses.Shorthandreferences
inregularexpressionsforspecifyingarangeofcharacters.
charactersetAnencodingsystemforrepresentingcharactersinbytes.
CHECKconstraintAruleonatablethatfiltersincomingdata.Onlydatathat
satisfiesthatrulewillbeacceptedbythetable.AlsoknownasaCHECK
integrityconstraint.
childAroworrecordthatisonelevelbelowanotherlevelinahierarchicaldata
relationship.Forexample,ifonetablecontains“orders”andanothercontainsthe
“lineitems”thateachordercontains,thenatablecontainingthose“lineitems”
wouldbesaidtobethechildtable.Achildtableisonethattypicallyhasa
foreignkeyrelationshipwithaparenttablesothatrowsintheparenttableare
onelevelhigherinthehierarchythantherowsinthechildtable.Seealso
orphan;parent.
clauseAsubsetwithinalargerconstruct,suchasaportionofastatementor
command.
CLOBCharacterLargeObject.Adatatypethatstoreslargeamountsof
characterdata,upto128terabytes.CLOBdatatypescanberolledbackor
committedaspartofanydatabasetransaction.
CoddThelastnameofDr.E.F.Codd,thepersoncreditedwithformingthe
originalideasthatledtothecreationofmodern-dayrelationaldatabase
programming.
columnAverticalspaceinadatabasetable.Columnshaveanameandadata
type.
commandAdirective.
COMMENTASQLstatementtoaddcommentstothedatadictionaryfor
databaseobjectsyouhavecreated.
commitTocausechangeswithinthecurrentsessiontobemadepermanent.
COMMITASQLstatementtosavedatatothedatabase.
conditionAnexpressionthatevaluatestoameaningfulresulttoindicatethe
nextcourseofaction.Conditionsareusedtodeterminewhetherastatementwill
takeaparticularactionornot;thedecisionhingesonwhetherthecondition
evaluatestotrueorfalse.
conditionalAsituationthatdependsontheevaluationofacondition.
connectEstablishausersessionwiththedatabase.
constantSeeliteral.
constraintAruledefininghowdataistobeprocessed.Atablecanhaveoneor
moreconstraintsthatmayrestrictittohavingcertainkindsofdataandrejecting
others.Seealsoreferentialintegrity.
conversionTheactoftransformingsomethingfromoneformtoanother.
ConversionfunctionsinSQLcanchangedatafromonedatatypetoanotherdata
type.
correlatedsubqueryAsubquerythatuses,aspartofitsexecutionlogic,data
fromanouterquery.
CREATEAreservedwordthatstartsoneofmanySQLstatements,usedto
createdatabaseobjectssuchastables,views,indexes,sequences,andsynonyms.
cross-joinSeeCartesianproduct.
datadictionaryAsetoftablesandviewsautomaticallymaintainedbythe
Oraclesystemfordocumentingthecharacteristicsandstatusofallobjectsinthe
database.
datafileAphysicalfileinthefilesystemforstoringdata,locatedeitherinthe
operatingsystemfilesystemorinanAutomatedStorageManagementdisk
group.
datatypeAsetofrulesdefiningasubsetofdata.
databaseAnorganizedcollectionofinformation.
databaseadministratorOftenabbreviatedDBA.Thejobofadministeringthe
database.TheDBAoftenistaskedwithinstallingthedatabasesoftwareand
configuringitforuse,performingbackups,generallymakingthedatabase
systemavailableforuse,maintainingoptimalperformance,andtakingstepsto
protectagainstlossofdata.
dateAcalendardate.
datetimeAnyofthesetofdatatypesDATE,TIMESTAMP,TIMESTAMP
WITHTIMEZONE,orTIMESTAMPWITHLOCALTIMEZONE.
daylightsavingtimeTimedefinedasa1-houroffsetfromstandardtimeto
providemoredaylightattheendoftheworkingdayduringseasonsinwhich
daylightislimited.
DBASeedatabaseadministrator.
DBMSDatabasemanagementsystem.
DDLDataDefinitionLanguage.AsubsetofSQL.ReferstothesetofSQL
statementsthatisusedtocreatedatabaseobjects,modifytheirstructure,and
removethemfromthedatabase.
defaultWhenusedinassociationwithaparameter,“default”isthevalueofthat
parameterwhennospecificvalueisassigned.Seeparameter.
DEFINEAreservedwordusedinpatternmatchingtospecifypatternvariables.
DELETEASQLstatementusedtoremovedatafromthedatabase.
deprecatedSaidofafeaturethatmaystillexistbutisnolongerofficially
supportedandwhoseuseisofficiallydiscouraged.
developerAnindividualengagedinthejobofcreatingapplications.
developmentTheactofcreatingapplications.
DMLDataManipulationLanguage.AsubsetofSQL.ReferstothesetofSQL
statementsusedtoqueryexistingdataindatabaseobjects,adddatatoexisting
databaseobjects,modifythatdata,andremovedatafromthedatabase.
DROPAreservedwordusedtostartoneofmanySQLstatements,allofwhich
areusedtoremovecertainexistingdatabaseobjectsfromthedatabase.
EllisonThelastnameofLarryEllison,founderofOracleCorporation,thefirst
companytoreleaseacommercialRDBMSproduct.
entityAnorganizedcollectionofattributesinadatamodel.Correspondstoa
table.Seealsoattribute;ERD.
equijoinAjointhatusesanequalityoperator(theequalsign)inthejoin
condition.
ERDEntity-relationshipdiagram.Adiagramthatshowsentitiesandhowthey
relatetoeachother.Seealsoentity.
escapecharacterAsinglecharacterthatcantakeonanalternativepurpose
separatefromitscharacter’sstandardmeaning.Forexample,asinglequotation
markisanescapecharacterwhenprecedinganothersinglequotationmarkina
textstringdelimitedbysinglequotessothatstringssuchas‘O’’Brian’willbe
correctlyinterpretedasO’Brianinthedatabase,ratherthanthetruncatedstring
‘O’followedbythecharactersBrian’—whichwouldbemeaninglessinanySQL
statementandwouldresultinasyntaxerror.
explicitcommitTheCOMMITstatement.
expressionAcombinationofliteralvalues,operators,variables,andfunctions,
withtheintentofcomputingavalueofaparticulardatatype.
externaltableASQLtablethatstorestablemetadatainthedatabasebutstores
thetable’sdataoutsideofthedatabase.
FIRSTAreservedwordthatreferstothefirstrowinasequenceofrows.
FLASHBACKASQLstatementusedtorestoreolderversionsofdatabase
objects.
flashbackoperationsAsetofoperationsusingundodatatosupportdata
recoveryandtheanalysisofhistoricaldataovertime.
foreignkeyAreferentialintegrityconstraintinatable.Aforeignkeyspecifies
oneormoreattributesinoneentitythatrelatetooneormoreattributesin
anotherentity.Valuesenteredintheforeignkey’sattributesmustalreadyexistin
thereferencedtable’scorrespondingattributes.Seealsoreferentialintegrity;
constraint.
functionAsetofcodethatperformsaparticulartaskandreturnsasingleresult.
ASQLfunctioncanbeusedwithinaSQLexpression.Afunctionisonetypeof
subprogram,alsoknownasaprogramunit.Thereisanotherformknownasa
procedure,whichisnotincludedontheexam.
GRANTASQLstatementusedtogivesystemprivilegesorobjectprivilegesto
auseraccount.
hierarchicalqueryAquerythatspecifiesmultiplelevelsofrelationship.
Typicallybuiltonaself-join.Notethatatypicaljoinoftwotablesinaparent-
childrelationshipcanbesaidtobeatwo-levelhierarchy;technicallythatwould
beaccurate.ButthetermhierarchicalqueryinOracleSQLisgenerally
understoodtoindicateaparticulartypeofquerybasedonadatamodelthatis
capableofsupportingmorethantwolevels.
IEEEInstituteofElectricalandElectronicsEngineers.Anonprofitorganization
withthemissiontoadvancetechnologyasitrelatestotheuseofelectricity.
implicitcommitAcommiteventotherthantheCOMMITstatement.The
executionofDDLcodewillresultinanimplicitcommit.
indexAdatabaseobjectthatcopiesasubsetofdatafromatable,presorted,and
intendedtosupportfasterqueryingontheindexedtable.
inlineviewAsubquerythatperformslikeaviewinsupportofasingleSQL
statement.
innerjoinAjoinoftwoormoretablesinwhichajoinconditionisspecified,
andtheresultconsistsexclusivelyofrowsinonetablethatmatchrowsinthe
othertableaccordingtothejoincondition.Ifarowinonetablehasnomatching
counterpartintheothertable,itisnotincludedintheresults.Seealsoouterjoin.
INSERTASQLstatementusedtostoredatainadatabasetable.
instanceOnesetofOraclebackgroundprocessesandmemorystructuresusedto
accessadatabase.
integrityconstraintSeeconstraint.
interpolationAmethodforcalculatingoneormorevalueswithinarangeof
knownvalues.
invisiblecolumnAninvisiblecolumninatableisnotvisible,butitsvaluesare
availabletobereferencedbySQLstatements,includingINSERTandSELECT
statements,andincludingVIEWobjectsbasedonthetable.
invisibleindexAnindexthatcanbemadeinvisibletotheOracleDatabase
QueryOptimizertoomittheinclusionofthatindexfromtheexecutionplanofa
givenSQLstatement.
joinTheactofconnectingrowsinonetablewithrowsinoneormoreother
tables,basedonsomecriteriathatdeterminehowthedatainthetablescorrelates
toeachother.
keyOneormoreattributes—orcolumns—usedinthedefinitionofanintegrity
constraint.Keysincludeprimarykeys,foreignkeys,anduniquekeys.
keywordAspecialwordusedinaSQLcommandorservingsomeotherspecial
purpose.Keywordsareoftenreservedwords,buttheyarenotnecessarily
reserved.
LASTAreservedwordthatreferstothelastrowinasequenceofrows.
linearReferstoastraight-linenumericpattern,asopposedtoexponential
patterns.Forexample,linearinterpolationconsidersknownvaluesinarange
andcomputesinterimvalues,assumingalinearrelationshipamongallvalues.
literalAfixeddatavalue.Alsocalledaconstant.
LOBLargeObject.Anyofanumberofdatatypesthatstorelargeamountsof
information.SeealsoBLOB;CLOB;NCLOB.
lowercaseThelettersofthealphabetinminisculeform,inotherwords,a,b,and
soon.
MERGEASQLstatementthatperformsacombinationofINSERT,UPDATE,
and/orDELETEstatementfunctionality.
metacharacteroperatorsUsedtodefinepatternsinregularexpressions.
metadataDataaboutdata.Forexample,isthe“accountnumber”atagiven
organizationanumericvalue,orisitanalphanumericvalue?Orperhaps
alphabetic?Metadatadescribesdatainhigh-levelterms.
multitableinsertASQLINSERTstatementthatisabletoaddrowsofdatato
oneormoretables.Multitableinsertscanbeconditionalorunconditional.
namespaceAvirtuallocationwithinthedatabaseinwhichnodatabaseobjects
maysharethesamename.Allnamesmustbeuniquewithinagivennamespace.
naturaljoinAjoininwhichthejoincriteriaareimpliedbasedoncommon
namesofcolumnsinthetablesbeingjoined.
NCLOBNationalCharacterSetLargeObject.Adatatypethatstoreslarge
amountsofcharacterdatainanationaldatabasecharacterset.Storesupto128
terabytes.NCLOBdatatypescanberolledbackorcommitted.
NEXTAreservedwordthatreferstothenextrowinasequenceofrows.
NLSNationalLanguageSupport.
NLSparametersVariablesthatcustomizethebehaviorofthedatabasein
accordancewithagivenlocale.NLS_SORTisanexample.
non-equijoinAjoinconditionthatusesoperatorsotherthantheequality
operatortospecifythejoincondition—suchasgreater-thanorless-than
operators.
normalizationAspecificseriesofprocessesintendedtosupportthedesignofa
databasetomaximizeefficiency.
NoSQLRefersto“notonlySQL”databases.
NULLUnknown.Theabsenceofinformation.
numberAdigit.
numericSaidofasetofdatatypesthatacceptnumberdata.
objectAniteminthedatabase.Objectshavepropertiesofstructureandsecurity.
objectprivilegeTherighttoperformaparticulartaskonaparticularobjectin
thedatabase.Seealsosystemprivilege.
OFFSETAreservedwordusedwithFETCHtolimitrowsreturnedbyaquery,
startingafterthenumberofrowsspecifiedbyOFFSET.
ONEROWPERMATCHReservedwordsusedwhenperforminganalytics
acrossmultiplerowsofdata.
operatorprecedenceTherulesdefiningtheorderinwhichoperatorswithinan
expressionareprocessed.
operatorsSymbolsthatperformtasksonvalueswithinanexpression.
optimizerAfeaturebuiltintotheOracledatabasethatevaluatesprocessing
withinthedatabasetoimprovespeedandoptimizerperformanceofSQL
statements.
ORA_ROWSCNAconservativeupperboundofthelatestcommittimeforthe
transactionthatlastchangedtherow.TheactualcommitSCNofthetransaction
canbesomewhatearlier.Seealsosystemchangenumber(SCN).
OracleTheleadingRDBMSproductonthemarkettoday.
OracleCorporationThefirstcompanytoproduceacommercialRDBMS
product.
orphanAchildrowinachildtableforwhichthereisnocorrespondingparent
rowinthecorrespondingparenttable.
outerjoinAjoinoftwoormoretablesinwhichajoinconditionisspecified,
andtheresultconsistsofrowsinonetablethatmatchrowsintheothertable
accordingtothejoincondition,aswellasrowsthatdonotmatch.Ifarowinone
tablehasnomatchingcounterpartintheothertable,itmaybeincludedinthe
results.Seealsoinnerjoin.
parameterAvariablethatispassedtoorfromafunctionorprocedure.
parentAroworrecordthatisonelevelaboveanotherlevelinahierarchical
datarelationship.Forexample,ifonetablecontains“orders”andanother
containsthe“lineitems”thateachordercontains,thenatablecontainingthose
“orders”wouldbesaidtobetheparenttable.Aparenttableisonethatis
referencedbyaforeignkeyinachildtablesothatrowsintheparenttableare
onelevelhigherinthehierarchythantherowsinthechildtable.Seealso
orphan;child.
parseToanalyzecodeforsyntacticaccuracy.SQLcodethatissubmittedfor
executionisparsedfirstandthenexecuteduponsuccessfulcompletionofthe
parsingprocess.
PARTITIONBYReservedwordsusedtospecifyasubsetofrowswithina
largerquery.
patternvariablesApatternvariableisanameyoucreateandassociatewitha
conditionyouspecify.Patternvariablesareusedtosearchformatchesby
comparingdataacrossmultiplerowsofdatawithcomparisonoperators.
POSIXPortableOperatingSystemInterface(forUnix).AsetofIEEEstandards
fordefiningstandardsandinteroperabilityonanumberofissues.
precedenceAlogicalprioritizationofasetofitems.
precisionPartofthedefinitionofanumericdatatype.Precisionspecifiesthe
numberofsignificantdigitsinthenumericvalue.Seealsoscale.
predicatesThesecompareoneexpressiontoanothertoproduceatrue,false,or
NULLresult.CanbecombinedwithBooleanoperatorsAND,OR,andNOT.
PREVAreservedwordthatreferstothepreviousrowinasequenceofrows.
primarykeyAuniquenon-NULLattributeinanentity,orauniquenon-NULL
columninatable.
privatesynonymAsynonymthatisnotaPUBLICsynonym.Thereisno
PRIVATEkeyword.
privilegeTherighttoperformataskinthedatabase.Seealsoobjectprivilege;
systemprivilege.
procedureAsetofcodethatperformsaparticulartask.Aproceduremayreturn
anywherefromzerotomultipleresults.ProcedurescannotbeusedwithinaSQL
expressionbutinsteadareofteninvokedinstatementsbythemselves.
productionProfessionaluse.Databaseapplicationsin“production”areactively
storingdataforanongoingorganization,asopposedtodatabaseapplications
thatareindevelopmentortesting.
projectionTheconceptofqueryingasubsetofcolumnsfromatable.
pseudocolumnsValuesthataredefinedautomaticallybytheOraclesystemfor
certainobjectsinthedatabase,suchastablesandsequences.Pseudocolumnscan
beselectedlikeacolumninatable.
PUBLICAspecialdatabaseuserautomaticallymaintainedbythedatabase.
PUBLICrepresentsallusersinthedatabase.GrantingprivilegestoPUBLIChas
theeffectofgrantingthemtoallusers.
PURGEASQLstatementtoremoveobjectsfromtherecyclebin.
queryASELECTstatement.Arequestofthedatabaseforsomeofthedata
containedwithinit.
RDBMSRelationaldatabasemanagementsystem.
readconsistencyTheabilityfordatainthedatabasetobereadandjoinedina
mannerthatisaccurate.Readconsistencyrepresentsaviewofdatathatis
“frozen”inaninstantoftime.Readconsistencybecomesimportantwhen
joiningtablesthatarebeingmodifiedinrealtimesothatasthedatabasequeries
onetableandthenanother,thecombinedrecordsreflectwhatwasintended.
recordAsetofdataelementsrelatedtoeachotherandrepresentinga
meaningfulcollectionofinformation.Onerowcanbearecord;joinedrows
mightalsobearecord.
recyclebinThestructureintheSQLdatabaseintowhichdroppedobjectsare
tracked.
redologsAsetofoperatingsystemfilesthatrecordallchangesmadetoa
database,whetherthosechangeshavebeencommittedornot.
referentialintegrityAconstraint,orrule,ononetable’scolumnthatrequires
anyvaluetobestoredinthatcolumntobealreadypresentinanotherparticular
table’scolumn.Seealsoforeignkey.
regularexpressionAlanguageofpatternmatching.Nottobeconfusedwith
expressions.Oracle’ssupportforregularexpressionsisconsistentwiththe
POSIXandUnicodestandards.
relationalHavingarelationorbeingrelated.Adatabaseissaidtoberelational
whenitisbuiltondataobjectsthatcanbejoinedtogetherbasedoncommon
criteriawithinandamongtheobjects.
RENAMEASQLstatementusedtochangethenameofcertainobjectsinthe
database.
reservedwordSpecialwordssetasideforspecialuseandnotavailablefor
applicationdevelopment.Youcannotusereservedwordsasthenamesof
databaseobjectsorvariables.
restorepointAmarkedpointintime,toberecordedforpossiblefuture
referenceinsupportofflashbackoperations.
REVOKEASQLstatementtoremovesystemprivilegesorobjectprivileges
thathavebeengrantedtoauseraccount.
roleAcollectionofoneormoreprivileges.
rollbackAnactionthatrestoresthedatabasetothemostrecentcommitwithin
thecurrentsession.
ROLLBACKASQLstatementusedtorestorethedatabasetoanearlierstate.
Cancelstheeffectsofatransactioninprogress.
rowOnesetofvaluesforthecolumnsofatable.
savepointAmarkedpointintime,toberecordedforpossiblefuturerollback.
SAVEPOINTASQLstatementthatmarksapointinasession.Futureusesof
theROLLBACKstatementmaychoosetorestorethedatabasetothepoint
markedbyaSAVEPOINTstatement.
scalarsubqueryAsubquerythatreturnsonecolumninonerowasitsoutput—
inotherwords,asinglevalue,asopposedtorowsofvaluesorcolumnsof
values.
scalePartofthedefinitionofanumericdatatype.Scalespecifieswhere
roundingwilloccurinthenumericdatatype.Seealsoprecision.
schemaAcollectionoftablesownedbyauseraccount.
SCNSeesystemchangenumber(SCN).
segmentAleveloflogicaldatabasestorage.
SELECTASQLstatementusedtoqueryoneormoredatabasetables.
selectionTheabilitytoqueryasubsetofrowsfromatable.
selectivityThedegreeofuniquenessofvaluesinacolumn.Ifallvaluesinthe
columnareidentical,selectivityissaidtobelow.Ifthevaluesareallunique,
selectivityissaidtobehigh.
self-joinAjointhatconnectsrowsinatablewithotherrowsinthesametable.
semijoinAquerythatreturnsrowsthatmatchanEXISTSsubquery.
sequenceAnumbergenerator.Adatabaseobject.
sessionAuserprocessinwhichtheuserinteractswiththedatabase.
setoperatorAnyoftheoperatorsUNION,UNIONALL,INTERSECT,or
MINUS.
SKIPAreservedwordusedinpatternmatching.Maybecombinedwithother
keywordstoformkeywordphrases,forexample,AFTERMATCHSKIPTO
NEXTROW.
SQLSeeStructuredQueryLanguage.
standarddeviationWhenusedwithnumericanalysis,standarddeviationisa
quantifieddegreeofvariationofindividualvalueswithinalargesetofvalues.
Theformulaforstandarddeviationtakesthesquarerootofthevariance.See
varianceformoreinformation.
standardtimeAlsoknownaswintertimezones.TimeasdefinedbyUTC.
statementAcommand.
stringAseriesofcharacters.
StructuredQueryLanguageAworldwidestandardlanguageforinteracting
withadatabase.
subqueryASELECTstatementcontainedwithinanother(outer)SELECT
statementsothatthedataofthesubqueryfeedsintotheprocessingoftheouter
query.
superaggregateAnaggregationofaggregatevalues.
synonymAnalias,oralternativename,forsomethinginthedatabase.A
synonymisitselfanobjectinthedatabase.
syntaxTherulesforformingastatement,acommand,orsomeotherlanguage
construct.
SYSAbuilt-inuseraccountwithDBAprivilegesthatcomeswithallOracle
installations.SYSownsthedatadictionary.
SYSTEMAbuilt-inuseraccountwithDBAprivilegesthatcomeswithall
Oracleinstallations.
systemchangenumber(SCN)Amarkerthatspecifiesacommittedversionof
thedatabaseataparticularpointintime.Eachcommittedtransactionisassigned
anSCN.Seealsotransaction.
systemprivilegeTherighttoperformaparticulartaskinthedatabase.Seealso
objectprivilege.
tableAstorageunitinthedatabasethatconsistsofcolumnsandrows.
tablespaceAmechanisminthedatabasethatishometooneormoretablesand
storesthatdatainoneormoredatafiles.
TCLTransactionControlLanguage.AsubsetofSQL.ReferstothesetofSQL
statementsthatisusedtocontrolauser’ssessioninwhichDMLstatementsare
used.TCLdetermineswhethertheresultsofaDMLstatementareallowedtobe
madepermanentorwhethertheyareundonefromthedatabase.
textCharacter-baseddata.
timezoneAregionoftheearththatusesuniformstandardtimeasanoffsetfrom
UTC.Therearecurrently24suchregionsdefinedintheearth,dividedroughly
bylongitudinallines.Alsoknownastimezoneregion.
timezonenameThenameofatimezoneregion.Examplesare
Pacific/AucklandandAmerica/Indianapolis.
timezoneoffsetAtimedifferencebetweenthelocaltimeandUTC.
timestampAvaluerepresentingthedateandtime.
transactionAseriesofoneormoreSQLstatementsexecutedbetweencommit
events.
TRUNCATEASQLstatementusedtoremovedatafromadatabasequicklybut
withoutrecoverability.
unconditionalWithoutrestriction.
undosegmentsSegmentsthataremaintainedautomaticallybythedatabaseto
supportrollbackoperations,toassurereadconsistency,andtootherwiserecover
fromlogicalcorruptions.
UnicodeAnindustrystandardthatattemptstocreateastandardizedencodingof
everycharacterofeverylanguageinexistence.
uniqueOneofakind.
uniqueidentifierAnunambiguousreferencetosomething,leavingnodoubt
whatisbeingreferenced.
UPDATEASQLstatementusedtomodifydatainadatabasetable.
uppercaseThelettersofthealphabetinmajusculeform,alsoknownascapital
letters,suchasA,B,andsoon.
useraccountAprocessthatprovidespassword-protectedaccesstoand
ownershipofasetofdatabaseobjectsandprivileges.
UTCCoordinatedUniversalTime.ThenewnameforGreenwichMeanTime.
Theuniversalstandardformeasuringtimeinternationally.UTCmeasurestime
asitexistsattheRoyalObservatoryofGreenwich,London.
variableAsmallunitofstorage,representedbyanameandadatatype,for
holdingvaluesthatcanbechanged.
varianceWhenusedwithnumericanalysis,variationquantifiesthedegreeof
variationofindividualvalueswithinalargesetofvalues.Varianceisdetermined
byaveragingasetofvalues,subtractingeachvaluefromtheaverage,squaring
eachdifference,andsummingtheresult.
viewAnamedquerythatisstoredinthedatabase.
windowWhenusedwithanalyticfunctions,awindowisagroupofrows
surroundingatargetrow.Analyticsmayperformoperationsonthetargetrowby
leveragingdatafromthroughoutthewindow.
wintertimezoneSeestandardtime.
INDEX
Pleasenotethatindexlinkspointtopagebeginningsfromtheprintedition.
Locationsareapproximateine-readers,andyoumayneedtopagedownoneor
moretimesafterclickingalinktogettotheindexedmaterial.
NUMBERS
1GL(first-generationlanguage),569
1NF(firstnormalform),23–25
2GL(second-generationlanguage),569
2NF(secondnormalform),23–25
3GL(third-generationlanguage),569
3NF(thirdnormalform),23–25
4GL(fourth-generationlanguage),569
A
ACCEPTcommand,substitutionvariables,178–180
accesscontrol.Seeuseraccesscontrol
ACCESSPARAMETERS,createexternaltable,89–90
ADD_MONTHSdatefunction,221
ADMINOPTIONclause,GRANTstatement,539–540
administrator.SeeDBA(databaseadministrator)
aggregate,defined,569
aggregatefunctions.Seealsogroupfunctions
commonlyused,281
defined,223
DENSE_RANK,292
FIRSTandLAST,292–293
GROUPBYrulesfor,296–298
manygroupfunctionsusedas,280
nestingrulesfor,300–303
RANK,290
aliases,569
ALLkeyword
conditionalmultitableINSERT,500–501
grantcertainprivilegestoallusers,541
unconditionalmultitableINSERT,499,501
ALLoperator
AVGfunction,288
comparison,multi-rowsubquery,371
COUNTfunction,285
ALL_prefix,datadictionaryviews,476–477,548
ALLPRIVILEGESoption,541,546–547
alphabetic,570
alphanumeric
inampersandsubstitution,175
characterdatatypesas,57
metadataas,570
ALTERSESSIONstatement,28,60,426
ALTERstatement
asDDLstatement,28
defined,570
implicitcommitwhenexecuting,124
ALTERSYSTEMstatement,28,60
ALTERTABLEstatement
createconstraints,66–67
createforeignkeys,75
FLASHBACKTABLElimitations,430–431
ALTERUSERstatement,534
ALTERVIEWstatement,402–403
AmericanNationalStandardsInstitute(ANSI),7,570
ampersandsubstitution
ACCEPTandPROMPTcommands,178–180
DEFINEcommand,177,180–182
restrict/sortoutputatruntimewith,173–177
reviewdrill,190–191
reviewself-testQ&A,195–196,199–200
SETandSHOWcommands,177–178
asSQL*Pluscapabilityonexam,148
UNDEFINE,177
VERIFYsystemvariable,178
analyticfunctions.Seealsoaggregatefunctions
defined,570
DENSE_RANK,291
FIRSTandLAST,292–293
groupfunctionsusedas,280
LAGandLEAD,226–230
OVER,PARTITIONBY,andORDERBY,224–226
overviewof,223
PERCENTILE_CONT,232–233
RANK,289–290
STDDEV,230–232
ANDoperator,Boolean
HAVINGclausewith,304–305
operatorprecedence,168–169
overviewof,164–167
ANSI(AmericanNationalStandardsInstitute),7,570
ANYcomparisonoperator,multi-rowsubquery,371
ANYkeyword,systemprivileges,538–539
ASkeyword,createviews,395
ASCreservedword,ascendingsortinORDERBY,151–152
asterisk,COUNTfunction,284–285
attributes
*reservedfornaming,50
buildingrelationaldatabase,20–21
createsimpletable,46–47
indatabasenormalization,23–24,32
defined,570
entity-relationshipdiagramwith,18–19
autonomous,subqueriesas,352
AVGfunction,287–288,295
B
batchscripts,180
BEFOREDROP,recoverdroppedtables,426
BETWEENreservedword,WHEREclause,170–171
BLOB(binarylargeobject)datatypes
constraintrestrictions,80
defined,570
overviewof,62
Boolean
defined,570
HAVINGclause,304–305
limitrowsretrievedbyquery,164–169
NOToperator,166–168
operatorprecedence,168–169
AND/ORoperators,164–166
built-in
datatypes,61
defined,570
functions,203
C
callingfunctions,SQLstatement,203
Cartesianproduct,8,570
CASCADEclause,107–109,534
CASCADECONSTRAINTS,droppedcolumns,82–84
CASEconditionalexpression,265
caseinsensitive
defined,570
table/objectnamesas,48
casesensitive
characterdatatypecomparisonsas,161
datadictionarynames,422
defined,571
objectnamesindatadictionaryas,422
passwordsas,534
quotednamesas,50
textsearchesas,162,206–207
case,strings
INITCAPmanaging,207–209
UPPERandLOWERmanaging,206–207
CASTconversionfunction,263–264
CEILnumericalfunction,215
CHARdatatype,57
characterclass(POSIXcharacterclass),571
characterdatatypes
listof,57
MINandMAXgroupfunctionsfor,286–287
ORDERBYandNULL,157
characterfunctions
CONCATand||,208–209
INITCAP,207–208
INSTR,212
LENGTH,211–212
LPAD,RPAD,209–210
LTRIMandRTRIM,210–211
overviewof,206
SOUNDEX,213–214
SUSTR,212–213
tasksperformedby,204
TRIM,211
UPPERandLOWER,206–207
characterset,defined,571
characters
defined,571
LIKEoperatorenablingwildcard,160,162–164
rulesfortable/objectnames,48
CHECKconstraint
createdattimeoftablecreation,77–78
datatyperestrictionson,79–80
defined,571
child,defined,571
childtables,recursivelytruncate,107–109
CHUNKparameter,BLOBdatatype,62
clause,defined,571
CLOB(characterlargeobject)datatype,62,80,571
Codd,Dr.E.F.,571
columnalias,154–155
columns
createmultipleindexesonsamesetof,423–424
CREATETABLEwith,53–54
datatypesavailablefor.Seedatatypes,columndefined,571
designatetableindexes,413
dropped,80–84
inDUALtable,206
groupingmultiple,298–299
insertrowsviadefaultcolumnlist,110–113
insertrowsviaenumeratedcolumnlist,113–116
inspectconstraintsindatadictionary,487–488
inspectindatadictionary,484–485,488
multiple-columnsubqueries,356
settoUNUSED,84–86
storageoftable,390
structureof,391
inUSER_TABLESview,476
visible/invisible,403–407
command,defined,571
COMMENTS
addtodatadictionary,480–482
defined,571
asDMLstatement,29
readindatadictionary,479–480
commit,defined,572
COMMITstatement
controltransactions,122
defined,30,572
explicitcommit,123–124
implicitcommit,124–125
andmultiplesessions,125–127
rulesforusingSAVEPOINT,129–131
comparisonoperators
Booleanlogic,164–169
ISNULLandISNOTNULL,171–172
LIKE,162–164
multi-rowsubqueries,370–372
single-rowsubqueries,369
WHEREclause,158–160
COMPILEstatement,402–403
compileviews,datadictionary,485–486
compositeindexes,418–420
compositeUNIQUEconstraint,71
CONCATand||characterfunction,208–209
concatenationoperators,209–210
condition,defined,572
conditional,defined,572
conditionalexpressions,SELECT
CASE,265
DECODE,266–267
defined,246
NULLIF,268–269
NVL,267–268
overviewof,264
review,269–271,273–275,277–278
conditionalmultitableINSERTstatements,499,500–502,505–513
connect,defined,572
CONNECTrole,551
CONNECTstatement,SQL*Plus,534–535
constant.Seeliterals
constraints
cannotbeaddedtoexternaltables,87
CHECK,77–78
createinALTERTABLE,66–67
createinCREATETABLE,53–54,63–66
create,modify,ordropwithALTERVIEW,402–403
indatadictionaryviews,487–488
datatyperestrictionson,79–80
defined,44,572
ONDELETEclause,76–77
FOREIGNKEY,73–77
fulllistof,78–79
implicitindexcreation,414–415
INSERTstatementand,115–116
multipleconstraints,78
naming,50
NOTNULL,67,69–71
preventsINSERT/UPDATE/DELETEonviews,398–399
PRIMARYKEY,71–73
recoverwithFLASHBACKTABLE,425–429
asschemaobjects,390–391
system-assignednames,51
UNIQUE,71
UPDATEstatementand,119–120
viewsand,397
CONSTRAINT_TYPEcolumn,datadictionaryviews,487–488
controltransactions
COMMIT,123–127
overviewof,122–123
ROLLBACK,127–129,132–133
SAVEPOINT,129–132
conversionfunctions
CAST,263–264
TO_CHAR,253–259
TO_DATE,259–260
defined,246
TO_DSINTERVAL,262
explicitandimplicit,247–249
marktimetorestoredata,433–434
TO_NUMBER,250–253
NUMTODSINTERVAL,263
NUMTOYMINTERVAL,262
overviewof,246–247
reviewcertificationsummary,269–270
reviewdrill,271
reviewself-testQ&A,272–278
TO_TIMESTAMP,260
TO_TIMESTAMP_TZ,261
TO_YMINTERVAL,261
CoordinatedUniversalTime(UTC),583,585
correlatedsubquery
defined,352,572
deleterowsinDELETE,362–363
Oracle1Z0-071vs.1X0-051examson,8
overviewof,356–357
querydatainSELECT,UPDATE,andDELETE,357–359
updaterowsinUPDATE,360–362
useof,355
COUNTfunction,283–285
CREATEANYTABLEprivilege,541
CREATEINDEXstatement
compositeindex,418
singlecolumnindex,416
specifyvisibilityofindex,422
CREATE,reservedword,572
CREATERESTOREPOINTstatement,434
CREATESEQUENCEstatement,408–409
CREATEstatement
asDDLstatement,28
asDMLstatement,29
implicitcommitsand,124
CREATETABLEstatement
createcompositePRIMARYKEYconstraint,73
createexternaltable,89–90
createforeignkeyswithALTERTABLEvs.,75–76
createin-lineconstraints,64–65
createout-of-lineconstraints,65–66
createPRIMARYKEYconstraint,72
createsimpletable,46–48
grantprivilegesonuserandtables,542–552
nametableorotherobject,48–53
overviewof,53–54
system-assignednames,51
usesubqueries,352–353,355
CREATEUSERstatement,533–534
CREATEUSERsystemprivilege,533–534
CREATEVIEWstatement,352–353,395–397
cross-join.SeeCartesianproduct
CURVALpseudocolumn,sequencegenerators,410–412
CYCLE,sequenceoption,409
D
DAB_SYS_PRIVSview,datadictionary,486
data
DMLstatementsworkswithobject,29–30
retrievefromtablewithSELECT,30–32
DataDefinitionLanguage(DDL).SeeDDL(DataDefinitionLanguage)
statementsdatadictionaryviews
addcommentsto,480–481
compileviews,485–486
defined,572
DICTIONARYview,481–482
dynamicperformanceviews,477–479
findcolumns,488
identifyuser’sownedobjects,482–484
inspectconstraints,487–488
inspecttablesandcolumns,484–485
manageobjectswith,473
Oracleexamson,8,11
queryvarious,474–475
readcomments,479–480
reviewcertificationsummary,489
reviewdrill,490–491
reviewself-testQ&A,492–496
forroles,551–552
storesinformationabouttables,390
structureof,475–477
viewprivilegesin,486–487,547–549
datafiles,573,584
data,manipulating
controltransactions,122–133
deleterowsfromtable,121–122
insertrowsintotable,109–116
reviewcertificationsummary,133–134
reviewdrill,135–137
reviewself-testQ&A,138–145
truncatedata,106–109
updaterowsintable,117–121
DataManipulationLanguage.SeeDML(DataManipulationLanguage)
statementsdatamodeling.SeeERD(entity-relationshipdiagram)
data,restrictingandsorting
ampersandsubstitution.Seeampersandsubstitutionlimitrows.See
rowsretrievedbyquery,limitoverviewof,148
reviewcertificationsummary,186–188
reviewdrill,189–191
reviewself-testQ&A,192–200
sortrowsretrievedbyquery.Seerowsretrievedbyquery,sortuseSQL
rowlimitingclause,182–186
datasets.Seelargedatasets
datatypes
comparing,160–162
conversionfunctionsfor.Seeconversionfunctionsdefined,573
explicitvs.implicitconversionof,247–249
noSQLBoolean,168
usingSELECTwithsetoperators,451–452
datatypes,column
built-invs.user-defined,63
character,57
date,59–61
insertrows/implicitconversionof,114–115
largeobject(LOB),61–62
numeric,58–59
overviewof,56
restrictionsonconstraints,79–80
database
defined,573
normalization,23–25
relationshipbetweenSQLand,25–27
date
defined,573
andNUMBERconstants,223
datedatatypes
formatmodels,254–260
functions.Seedatefunctions
listof,59–61
MEDIANgroupfunctionfor,288–289
MINandMAXgroupfunctionsfor,286–287
ORDERBYandNULLwith,157
datefunctions
ADD_MONTHS,221
DATEandNUMBERconstants,223
LAST_DAY,220
MONTHS_BETWEEN,221–222
NEXT_DAY,220
NUMTODSINTERVAL,222
NUMTOYMINTERVAL,222
ROUND—Date,218–219
SYSDATE,217
tasksperformedby,204–205
TRUNC—Date,219–220
datetime,573.Seealsodatedatatypesdatetimeconversionfunctions
TO_DSINTERVAL,262
NUMTOYMINTERVAL,262
TO_TIMESTAMP,260
TO_TIMESTAMP_TZ,261
TO_YMINTERVAL,261
daylightsavingtime
defined,573
TIMEZONE_HOURfieldforchangesin,59
TZDtimezonewith,256
DBA(databaseadministrator)
controlovertablespaces,535
defined,573
exclusiveaccesstoSYSaccount,475
GRANTALLPRIVILEGESoption,540
GRANTWITHADMINOPTION,539–540
role,551
tryingcodesamplesinthisbookand,530
DBA_prefix,datadictionaryviews,476–477
DBA_RECYCLEBIN,DROPTABLE,426
DBA_ROLE_PRIVS,datadictionaryview,487,552
DBA_ROLES,datadictionaryview,552
DBA_SYS_PRIVS,datadictionaryview,487,548–549,552
DBA_TABLES,datadictionaryview,476
DBA_TAB_PRIVS,datadictionaryview,487,548–549,552
DBMS(databasemanagementsystem),defined,573
DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBERfunction,SCN,
431,434
DDL(DataDefinitionLanguage)statements
automaticallyupdatesdatadictionary,475
defined,573
implicitcommitswhenexecuting,124
impliedcommitscausedby,30
managetables.Seetables,create/manageoverviewof,28–29
DECODEconditionalexpression,266–267
defaultcolumnlist,110–113
default,defined,573
DEFINE,reservedword,177,180–182,573
DELETEstatement
acceptssubqueries,352
COMMIT,multiplesessionsand,125–127
defined,573
deleterowsfromtable,121–122
deleterowswithcorrelatedsubquery,362–363
asDMLstatement,29
executeonviews,396–400
indexmaintainedwithevery.Seeindexesquerydataviacorrelated
subqueriesin,355,357–359
removerows,106
TRUNCATETABLEvs.,106–107
DELETE_RULEcolumn,datadictionaryview,487–488
DENSE_RANKfunction,291–292
dependentobjects,andtablerecovery,428
dependentprivileges,547
deprecated,defined,573
DESCstatement,SQL*Plus,7,54–56,152
DESCRIBEstatement
describeDICTIONARYview,481
describeviewwith,391,394
reviewtablestructurewith,54–56
descriptivenames,databaseobjects,51
developer,defined,573
development,defined,574
DIRECTORYobjects,87–90
DISABLENOVALIDATEstate,constraints,402
DISTINCToperator
withAVGgroupfunction,288
avoidLOBsusedwith,59
withCOUNTgroupfunction,285
preventsINSERT/UPDATE/DELETEonviews,399
DML(DataManipulationLanguage)statements
acceptssubqueries.Seesubqueries
defined,574
SQLstatementsforming,122
workswithexistingdatabaseobjects,29–30
workswithTCLstatements,30
documentation
additional,16
usingthisbookforexam,16
doublequotations
columnaliasoptions,155
quotednames,50–51
DROPCOLUMNstatement,80–85
DROPINDEXstatement,420–421
DROPRESTOREPOINTstatement,434
DROPstatement
defined,28,574
implicitcommitswith,124
restoredroppedtableBEFOREDROP,547
DROPTABLEstatement,426
DROPUNUSEDCOLUMNSstatement,86
DROPUSERstatement,534
DUALtable,206
DUMMYcolumn,DUALtable,206
dynamicperformancesynonyms
V$DATABASE,479
V$INSTANCE,479
V$OBJECT_USAGE,479
V$PARAMETER,479
V$RESERVED_WORDS,479
V$SESSION,479
V$TIMEZONE_NAMES,479
dynamicperformanceviews,477–479
E
elements,numericformatmodel,250–253
Ellison,Larry,574
ELSEclause,conditionalmultitableINSERT,499–501,506–508
ENDkeyword,CASEconditionalexpression,265
entity,574.Seealsoattributes;ERD(entity-relationshipdiagram)entity-
relationshipdiagram.SeeERD(entity-relationshipdiagram)enumeratedcolumn
list,113–116
equalssign,identifyrowswithWHEREclause,158–159
equijoins
defined,574
innerjoins,322–325
multitablejoins,330–331
naturaljoins,327–329
outerjoins,335–339
vs.non-equijoins,321
ERD(entity-relationshipdiagram)
anddatamodeling,17–20
databasenormalization,23–25
defined,574
many-to-manyrelationships,22–23
overviewof,16
relationaldatabasesand,20–21
escapecharacter,singlequote,207–208,252,570
examoverview
OracleSQLvs.ANSISQL,7
OracleSQLvs.OracleSQL*Plus,7
passingscore,11–12
SQLFundamentalsIvs.SQLCertifiedAssociate,7–12
studymaterials,13–16
subjectareas,12–13
testlogistics,4–7
EXISTSoperator,subqueries,363
EXPLAINPLAN,indexdesign,418
explicitcommit
defined,574
overviewof,123–127
rulesforSAVEPOINT,131
explicitconversion,247–249
expressions
attachtoconstraints,77–78
conditional.Seeconditionalexpressions,SELECT
conditionalmultitableINSERT,500
defined,574
dynamic,353
evaluatemultiple,164–169
GROUPBYclauserulesfor,296–297
INSERTstatementand,109
numberfunctionsin,204
ORDERBYclausefor,153–154,300
placecolumnaliasesafter,154
subqueriesasalternativeto,355
UPDATEstatementand,118–119
useSELECTwithsetoperators,451
WHEREclauseand,158–160
externaltables
create,89–90
defined,574
DIRECTORYobjects,87–88
Oracleutilitiesfor,88
restrictionson,86
workingwith,91
F
failedexecutionattempt,368
FETCHclause,SQLrowlimitingclause,182–186
fields,datedatatype,59–61
FIRSTfunction,292–293
first-generationlanguage(1GL),569
FIRSTkeyword
conditionalmultitableINSERT,500–501
defined,574
FETCH,182
firstnormalform(1NF),23–25
flashbackoperations
defined,575
dependentobjects,428
markingtime,431–434
overviewof,424–425
recoverdatainexistingtables,428–431
recoverdroppedtables,425–426
recyclebin,427
reviewcertificationsummary,436
reviewdrill,439
reviewself-testQ&A,444–445,448
statementexecution,428–429
FLASHBACKstatement
asDDLstatement,28
defined,575
FLASHBACKTABLEstatement
executionof,428
limitations,430–431
recoverdatainexistingtables,429–431
recoverdroppedtables,425–429
restoredataviaTIMESTAMP,432
restoredroppedtable/grantedprivileges,547
FLOORnumericalfunction,215
foreignkey
defined,575
jointabletoitselfusingself-joins,334
notrequiredtojointables,332
FOREIGNKEYconstraint,73–77,79–80
formatmodels
datedatatypes,254–260
numeric,250–253
fourth-generationlanguage(4GL),569
FROMclause
definedynamicviewswithsubqueries,353
GRANTstatementwith,538
inlineviewsreplacing,400–402
innerjoins,323
jointabletoitselfwithself-joins,334
multitablejoins,330–331
preventsINSERT/UPDATE/DELETEonviews,399
WHEREclausealwaysfollows,158–159
FULLOUTERJOIN,338–339
functions
analytical.Seeanalyticfunctions
applyconditionalexpressions,264–269
built-invs.user-defined,203
conversion.Seeconversionfunctions
defined,575
group.Seegroupfunctions
single-rowfunctions.Seesingle-row(scalar)functions
G
GRANTstatement
WITHADMINOPTION,539–540
ALLPRIVILEGES,540–541,546–547
asDDLstatement,28
defined,575
WITHGRANTOPTION,545–546
implicitcommitswhenexecuting,124
privilegesontablesanduser,542–552
privilegesvs.roles,552–553
PUBLICwith,541
roles,550–552
syntaxforsystemprivileges,536–538
GRANTEEcolumn,datadictionaryviews,549
GROUPBYclause
avoidusingLOBswith,59–61
groupdatawith,293–298
include/excludegroupedrows,303–304
multiplecolumns,298–299
nestingfunctions,300–303
ORDERBYclauseand,299–300
preventsINSERT/UPDATE/DELETEonviews,398
groupfunctions
asaggregateandanalyticfunctions,280
AVG,287–288
COUNT,283–285
DENSE_RANK,291–292
FIRST,LAST,292–293
groupdata.SeeGROUPBYclause
HAVINGclause,303–305
MEDIAN,288–289
MIN,MAX,286–287
others,292–293
RANK,289–290
reviewcertificationsummary,305–306
reviewdrill,307–308
reviewself-testQ&A,307–308
SUM,285–286
useof,280–283
GV_$prefix,dynamicperformanceviews,477–479
GV_$prefix,globaldynamicperformanceviews,477
GV$prefix,globaldynamicperformanceviews,477
H
HAVINGclause,include/excludegroupedrows,303–305
hierarchicalquery,defined,575
I
ID(identification),forexamregistration,4–5
IEEE(InstituteofElectricalandElectronicsEngineers),575
IF-THEN-ELSE,DECODEfunction,267
implicitcommit
COMMIT,multiplesessionsand,125–127
defined,123,575
invokedbyrestoreddata,430
overviewof,124–125
rulesforSAVEPOINT,131
implicitconversion,114–115,247–249
implicitindexcreation,414–415
in-lineconstraints,64–65,79
INoperator
multi-rowsubquery,371
single-rowsubquery,369
WHEREclause,169–170
INCREMENTBYinteger,sequences,408
INDEXobject,defined,43
indexes
alternativesonsamecolumnset,423–424
columns,421–422
defined,575
dropping,420–421
implicitcreationof,414–415
maintenance,418
OracleDatabaseoptimizerand,414
overviewof,412–413
recoverwithFLASHBACKTABLE,425–429
reviewcertificationsummary,436
reviewdrill,437–439
reviewself-testQ&A,443–444,448
asschemaobjects,391–392
singlecolumn,416–420
skipscanningincompositeindexes,419–420
unique,420
usage,416–418
INITCAPcharacterfunction,mixed-casestrings,207–208
inlineviews
createwithsubqueries,355
defined,575
reasonstouse,402
innerfunctions,234
innerjoins
defined,576
naturaljoinsas,327–329
oldersyntaxfor,324–325
overviewof,322–324
USINGkeywordfor,329–330
vs.outerjoins,321
INSERTstatements
addrowstotable,109
COMMIT,multiplesessionsand,125–127
conditionalmultitable,499–502,505–513
constraintsand,115–116
defined,576
asDMLstatement,29
executionerrorsin,116
indexesfor.Seeindexes
insertrowsviadefaultcolumnlist,110–113
insertrowsviaenumeratedcolumnlist,113–116
invisiblecolumnsandviews,404–407
permanentchangeswithCOMMIT,124
permanentresultswithGRANT,544
insingleMERGEstatement,514–517
subqueriesin,352–353,355
UPDATEvs.,121
forviews,396–400
instance,defined,576
InstituteofElectricalandElectronicsEngineers(IEEE),575
INSTRcharacterfunction,212
integrityconstraint.Seeconstraints
interpolation
defined,576
MEDIANperformslinear,289
PERCENTILE_CONTforlinear,232,235
INTERSECToperator
combinesmultiplequeriesintosinglequery,456–457
ensureSELECTcombineswith,451–452
inmultiplecombinations,458–459
assetoperator,450–451
INTERVALDAYTOSECONDdatatype,262
INTERVALYEARTOMONTHdatatype,61,183–186,261–262
INTERVALYEARTOSECONDdatatype,263
INTOclause
conditionalmultitableINSERT,499–501,505–511
mergerowsintotable,514–516
unconditionalmultitableINSERT,499,503–505
invalidviews,recompile,402–403
invisibleindexes,421–422
INVISIBLEkeyword,CREATEINDEX,422
invocationoffunctions,SQLstatements,203
ISNOTNULLoperator,WHEREclause,171–172
ISNULLoperator,WHEREclause,171–172
J
joins
avoidusingLOBswith,59
defined,320,576
displaydatafrommultipletables,320
ondynamicperformanceviews,478–479
impactonviews,403–407
inner,322–325
multitable,330–331
natural,327–329
non-equijoins,331–333
outer,335–339
reviewcertificationsummary,339
reviewdrill,340–341
reviewself-testQ&A,342–350
self-joins,333–335
tablealiasesusedby,325–327
typesof,320–321
USINGkeyword,329–330
K
key,defined,576
keyword,defined,576
L
LAGanalyticalfunction,226–230
largedatasets
featuresofmultitableINSERTs,498–502
mergerowsintotable,513–517
overviewof,498
performpivotwithmultitableINSERT,512–513
reviewcertificationsummary,518
reviewdrill,519–520
reviewself-testQ&A,521–527
useconditionalINSERTs,505–511
useunconditionalINSERTs,502–505
largeobjectdatatypes.SeeLOB(largeobject)datatypes
LASTfunction,292–293
LASTreservedword,576
LAST_DAYdatefunction,220
LEADanalyticalfunction,226–230
LEFTOUTERJOIN,338–339
LENGTHcharacterfunction,stringlength,211–212
LIKEcomparisonoperator,162–164,369
limitrows.Seerowsretrievedbyquery,limit
linear,defined,576
list
ofconstraintsintablecreation,78–79
ofdatabaseobjects,42–43
literals
alsocalledconstants,577
defined,577
LOB(largeobject)datatypes
cannotcreateindexoncolumnsof,413
convertLONGdatatypesto,62
defined,577
listof,61–62
restrictionsforexternaltables,87
logintouseraccount,withCONNECT,534
LONGdatatypes,convert,62
LOWERcharacterfunction,206–207
LPADcharacterfunction,209–210
LTRIMcharacterfunction,210–211
M
many-to-manyrelationships,23–24
marktime,restoredata,431–434
MAXfunction,286–287
MAX_STRING_SIZEparameter,VARCHAR2datatype,57
MAXVALUEinteger,sequences,408–409
MEDIANfunction,288–289
MERGEstatement
defined,577
asDMLstatement,29
mergerowsintotable,514–517
subqueriesacceptedby,352
metacharacteroperators,577
metadata,474–475,577
MINfunction,286–287
MINUSoperator
combinesmultiplequeriesintosinglequery,457–458
ensureSELECTcombineswith,451–452
inmultiplecombinations,458–459
assetoperator,450–451
MINVALUEinteger,sequences,408–409
mobilephone,inexamprocess,4,7
MODnumericalfunction,217
MONTHS_BETWEENdatefunction,221–222
multiple-choicequestions,onexam,5–6
multiple-columnsubqueries,356
multiplecolumns,grouping,298–299
multipleconstraints,65,78
multiple-rowsubqueries,356,369–372
multiplesessions,COMMITand,125–127
multirowfunctions.Seeaggregatefunctions
multitableINSERTs
defined,577
featuresof,498–502
pivotingwith,512–513
review,518–519
reviewself-testQ&A,521–524,526–527
useconditionalINSERTs,505–511
useunconditionalINSERTs,502–505
multitablejoins,330–331
N
namespaces
defined,577
namepriorityin,544
rolesand,551
understanding,51–53
naming
columns,inINSERTstatement,113
constraints,fromtablecreation,63–66
datadictionaryobjects,422
datadictionaryviewprefixes,476–477
retrieveddroppedobjects,428
SAVEPOINTstatements,130
schema,45
subqueries,usingWITH,364–365
tablesorotherobjects,48
views,353,395
NationalLanguageSupport(NLS),250–253,577
naturaljoins,327–329,577
NCLOB(NationalCharacterSetLargeObject)datatype,62,577
nestedfunctions
conversionfunctionsas,260
overviewof,234
andpatterns,234–235
rulesfor,300–303
singleaggregatefunctions,288
nestedinlineviews,400
nestedqueries,355
NEXTkeyword,182,577
NEXTVALpseudocolumn,sequencegenerators,410–412
NLS(NationalLanguageSupport),250–253,577
NLS_DATE_FORMAT,datedatatype,60
NOCYCLE,sequences,409
NOMAXVALUEinteger,sequences,408–409
NOMINVALUEinteger,sequences,408–409
non-equijoin,321,336–339,578
non-scalar(oraggregate)functions,234–235
nonschemaobjects,42–43,45–46
normalization,23–25,578
NoSQL,defined,578
NOTEXISTSoperator,subqueries,363
NOTINoperator,multi-rowsubqueries,371–372
NOTNULLconstraint
conceptofNULL,69–71
createdattimeoftablecreation,65,69
datatyperestrictionson,79–80
preventsINSERT/UPDATE/DELETEonviews,398
inPRIMARYKEYconstraint,71
reviewof,79
syntaxvariationin,67–68
NOToperator
withBETWEENinWHEREclause,171
HAVINGclauseusing,304–305
mayprecedeINoperator,170
operatorprecedence,168–169
overviewof,166–168
NOToperator,multi-rowsubqueries,371
NULL
conceptof,69–70
COUNTfunctionand,284–285
DECODEconditionalexpression,266–267
defined,578
ISNULL,ISNOTNULL,171–172
MEDIANfunctionignores,288
multi-rowsubqueriesand,371–372
NOTNULLconstraintand,70–71
NULLIFconditionalexpression,267–268
NVLconditionalexpression,267–268
ORDERBYclauseand,157
resultsofSELECTwithORDERBY,150
UNIQUEallowsadditionof,71
NULLIFconditionalexpression,267–268
NUMBERdatatype,58–59
numbers,392,578
numericdatatypes
AVGfunction,287–288
defined,578
formatmodelsfor,250–253
listof,58–59
MEDIANgroupfunctionfor,288–289
MINandMAXgroupfunctionsfor,286–287
RANKgroupfunctionfor,289–290
numericliterals,223
numericalfunctions
CEIL,215
FLOOR,215
MOD,217
overviewof,215
REMAINDER,216
ROUND-Number,215
tasksperformedby,204
TRUNC-Number,216
NUMTODSINTERVALfunction,222,263
NUMTOYMINTERVALfunction,222,262
NVLconditionalexpression,267–268
O
objectprivileges
assigntoroles,550–552
create/grant/testontable,542–544
defined,531,578
dependentprivileges,547
grant/revokewithALLPRIVILEGES,546–547
grantroles,549–552
granttoPUBLIC,451
privilegesvs.roles,552–553
revokewithREVOKE,546
viewindatadictionary,549
vs.systemprivileges,530–531,533
objects
categorizemain,42–46
DDLstatementsbuild,28
defined,578
dependent,428
DMLstatementsworkwithdatain,29–30
identifyuser-owneddatadictionary,482–484
manage.Seedatadictionaryviews
naming,48–53
schema,2
OFFSET,reservedword,186,578
ONcondition,mergerowsintotable,514–516
ONDELETECASCADEclause,recursivelytruncatechildtables,107–109
ONDELETEclause,FOREIGNKEYconstraint,76–77
ONEROWPERMATCH,578
one-to-manyrelationships,databasenormalization,23–24
ONLYkeyword,FETCH,183
operatorprecedence,168–169,458,578
operators
Boolean,164–169
comparison.Seecomparisonoperators
defined,578
optimizer,578
OPTIMIZER_USER_INVISIBLE_INDEXESparameter,421
ORoperator,Boolean,164–166,168–169,304–305
ORREPLACEkeywords,88,395–397
OracleCorporation,defined,579
OracleDatabaseoptimizer
overviewof,414
single-columnindexes,416–420
SQLstatementprocessing,413
structuringSQLstatementforindex,417
visibleandinvisibleindexes,421–422
OracleDatabaseSQLCertifiedAssociate,4–5,7–12
OracleDatabaseSQLLanguageReferenceManual.SeeSQLLanguage
ReferenceManual
Oracle,defined,578
Oracleexam.Seeexamoverview
OracleSQL*Plus,OracleSQLvs.,7
ORACLE_LOADER,externaltables,89–90
ORA_ROWSCNpseudocolumn,432,578
ORDERanalyticalfunction,224–226
ORDERBYclause
analyticfunctionsand,223
avoidLOBswith,59–61
columnaliasesand,154–155
controlorderofrowsreturned,459–461
expressionsin,153–154
withinlineviews,401
innerjoinsand,323
withLAGandLEADfunctions,227,229
andNULL,157
aspartofSELECT,157–158
referencebyname,149–151
restrictuseofGROUPBYinSELECT,299–300
sortcolumns,156–157
sortdirectionwithASCandDESC,151–152
sortrowsretrievedbySELECT,157–158
WITHTIEStiedto,184–186
userowlimitingclauseafter,182–186
useSELECTwithsetoperators,452
ORGANIZATIONEXTERNAL,externaltables,89–90
orphan,defined,579
out-of-lineconstraints,65–66,79
outerjoins
defined,579
FULLOUTERJOINS,337–338
LEFTOUTERJOINS,336
naturaljoinsas,327–329
Oraclesyntax,338–339
RIGHTOUTERJOINS,336–337
self-joins,334
USINGkeywordfor,329–330
vs.innerjoins,321
OVERanalyticalfunction,224–229
overloadedfunctions,253
OWNERview,datadictionary,476
P
parameters
characterfunctions.Seecharacterfunctionsconditionalexpressions.See
conditionalexpressions,SELECT
conversionfunctions.Seeconversionfunctionsdatefunctions.Seedate
functions
defined,579
function,202
groupfunctions.Seegroupfunctions
NLS,250–253
numericalfunctions,215–217
parent
defined,579
queries,insubqueries,352–353
parse,defined,579
PARTITIONBY
defined,579
usedwithanalyticversionofRANK,289–290
usedwithanalytics,225–226
usedwithLAGandLEADfunctions,227–228
passwords
ascasesensitive,534
createuseraccounts,533–534
establishnewuseraccounts,535–536
patternvariables,defined,579
patterns,andnestedfunctions,234–235
PERCENTkeyword,FETCH,182
percentsign(%)wildcardsymbol,LIKEcomparisonoperator,162–164
PERCENTILE_CONTanalyticalfunction,232–233
performance,dynamicviewsof,477–479
performancetuning
benefitsofindexes,413
designindexesfor,418
withOracleDatabaseoptimizer,417–418
persistentstore,asrelationaldatabase,25
PIVOTkeyword,513
pivotingdata,conditionalmultitableINSERT,512–513
POSIX(PortableOperatingSystemInterfaceforUnix),571,579
precedence
Booleanoperator,168–169
defined,579
operator,578
setoperatorshaveequal,458
precision
defined,580
NUMBERdatatype,58–59
predicates,defined,580
prefixes
datadictionaryviews,476–477
dynamicperformanceviews,477–479
namingobjects,51
schema,544–545
PREV,reservedword,580
PRIMARYKEYconstraint
composite,72–73
createdattimeoftablecreation,64–66,71–73
datatyperestrictionson,79–80
FOREIGNKEYconstraintrequiring,73–75
implicitindexcreation,414–415
requiredataforcolumnusing,70–71
reviewof,79
primarykey,defined,580
PRIMARYKEY,sequencesand,392
privatesynonyms,46,580
privileges
categoriesof,530–531
checkingviewstoinspect,486–487
defined,580
dependent,547
object.Seeobjectprivileges
rolesas.Seeroles
rolesvs.,552–553
system.Seesystemprivileges
problems,solvedbysubqueries,354–355
procedure,defined,580
production,defined,580
projection,defined,580
PROMPTcommand,ampersandsubstitution,178–180
pseudocolumns
defined,580
inlineviewsandROWNUM,401
sequencegenerators,410–412
PUBLICdatabaseuser
defined,580
GRANTandREVOKEprivilegeson,541
PUBLICSYNONYMownedby,544
PUBLICSYNONYM
createforsimpletable,46
fordynamicperformanceviews,478–479
overviewof,45–46
renamedatadictionaryobjectsvia,475–477
schemaprefixes,544–545
PURGEstatement
defined,580
asDMLstatement,29
removeallobjectsinschema’srecyclebin,429
removelatestversionofitemfromrecyclebin,428
removeobjectsfromrecyclebin,427
Q
queries
combineinlineviewswithcomplex,401
combinemultiple,withsetoperator,452–455
combineviewswithcomplex,396
defined,581
ondynamicperformanceviews,478–479
hierarchical,575
indexesimproveperformanceof,412
limitrowsretrievedby.Seerowsretrievedbyquery,limitasSELECT
statements,352
solvingwithsubqueries.Seesubqueriesquestions,howtoprogress
throughexam,5–6
quotednames,tables/otherobjects,50–51
quotes,ampersandvariableand,175–176
R
RANKfunction,289–290
RAWdatatypes,413
RDBMS(relationaldatabasemanagementsystem)
captureactualdatawith,20–21
ERDand.SeeERD(entity-relationshipdiagram)andjoins,320
relationshipbetweenSQLand,25–27
readconsistency
defined,581
ofdynamicperformanceviews,478–479,489
undosegmentsassuring,585
recompileviews,datadictionary,485–486
record,defined,581
recovery
ofdatawithinexistingtablesovertime,429–431
ofdroppedtables,425–429
identifypointatwhichtorestoredata,431–434
recyclebin
defined,581
DROPTABLEand,426
PURGE,428,429
recoverdependentobjects,428
redologs,defined,581
referencebyposition,sortcolumns,156
references,exam.SeeSQLLanguageReferenceManual
referentialintegrity
defined,581
FOREIGNKEYconstraintensuring,73
registrationprocess,exam,4–5
regularexpressions,defined,581
relationaldatabasemanagementsystem.SeeRDBMS(relationaldatabase
managementsystem)relationship,betweendatabaseandSQL,25–27
REMAINDERnumericalfunction,216
RENAMEstatement
asDDLstatement,28
defined,581
RENAMETOkeywords,recoverdroppedtable,426
reservedwords
defined,582
keywordsoftenas,576
namesfordatabaseobjectscannotbe,48
quotednamesmayinclude,50
RESOURCErole,551
restorepoint
create,434
defined,582
recoverdroppedtables,425
restrict
data.Seedata,restrictingandsortingoutputatruntime.Seeampersand
substitutionREVOKEALLstatement,547
REVOKEstatement
ALLPRIVILEGESoption,547
basicsyntaxforsystemprivileges,538
asDDLstatement,28
defined,582
doesnotcascadeforsystemprivileges,540
implicitcommitswhenexecuting,124
onprivileges,546
privilegesvs.roles,553
onPUBLICaccount,541–542
RIGHTOUTERJOIN,338–339
ROLEobjects,defined,44,46
ROLE_ROLE_PRIVS,datadictionaryview,552
roles
defined,531,582
grant,550–552
grantprivilegesindirectlyas,548
privilegesvs.,552–553
ROLE_SYS_PRIVSview,datadictionary,487
ROLE_TAB_PRIVSview,datadictionary,487
rollback
defined,582
FLASHBACKTABLEstatementand,426
ROLLBACKstatement
controltransactionswith,127–129
defined,30,122,582
rollbacktonamedSAVEPOINT,132–133
rulesforusingSAVEPOINT,129–131
undochangesbeforeexecutingCOMMITvia,123
ROUNDdatefunction,218–219,220
ROUND-numberfunction,215
ROWNUMpseudocolumn,401
rows
controlorderofreturned,459–461
defined,582
findSCNforgiven,432
include/excludegrouped,303–305
insertintotables,109–116
mergeintotable,514–517
multiple-rowsubqueriesand,356,369–372
sequencessupportaddingtable,392
single-rowsubqueriesand,356,365–369
SQLlimitingclausefor,182–186
subqueriesreturncolumnsin,352
useSQLrowlimitingclause,182–186
rowsretrievedbyquery,limit
additionalconcepts,172–173
Booleanlogic,164–169
overviewof,157–158
WHEREclause,158–164,169–172
rowsretrievedbyquery,sort
ASCandDESC,151–152
columnalias,154–155
combinations,156–157
expressions,153–154
ORDERBYandNULL,157
overviewof,148–149
referencebyname,149–151
referencebyposition,156
RPADcharacterfunction,209–210
RTRIMcharacterfunction,210–211
rules
columnalias,154–155
combineSELECTwithsetoperators,451–452
createviews,395
datatypecomparisons,161
databasenormalization,23–25
GROUPBYclause,296–297
namedatabaseobjects,48–50
nestingfunctions,300–303
SAVEPOINT,130–131
runtime
valueofsubstitutionvariableat,173–177
violatedconstraintscausingerrorsat,115–116
S
savepoint,defined,582
SAVEPOINTstatement
controltransactionswith,129–132
defined,30,122,582
ROLLBACKtoparticularnamed,132–133
scalarfunctions.Seesingle-row(scalar)functions
scalarsubquery,353,357,582
scale,58–59,582
schema
create/useindexes.Seeindexes
create/usesequences,409–412
createviews.Seeviews
defined,582
flashbackand.Seeflashbackoperationsfunctionalityof,390–392
listof,42–43
nonschemaobjectsvs.,45–46
overviewof,44–45
prefixes,544–545
reviewcertificationsummary,435–436
reviewdrill,437–439
reviewself-testQ&A,440–448
SCN(systemchangenumber)
conversionfunctions,433–434
defined,582,584
identifyrestorepointvia,434
markpointatwhichtorestoredata,431–434
recoverdroppedtablesvia,425
restoredatausingTIMESTAMPvs.,432
SAVEPOINTnameassociatedwith,130
SCN_TO_TIMESTAMPfunction,433
SEARCH_CONDITIONcolumn,datadictionaryview,488
second-generationlanguage(2GL),569
secondnormalform(2NF),23–25
segments
defined,582
undo,425,585
SELECTCURRENT_SCNFROMV$DATABASE,431
SELECTstatements
applyconditionalexpressions,264–269
buildtoretrievedatafromtable,30–32
WITHclause,364–365
combinewithsetoperators,451–459
controlorderofrowsreturned,459–461
defined,583
DML,29–30
functionsin.Seesingle-row(scalar)functionsGROUPBYuniqueto.
SeeGROUPBYclauseidentifyrowswithWHEREclause,158–159
madeintosubqueries.Seesubqueries
querydataviacorrelatedsubqueriesin,355,357–359
restrictingandsortingdatawith.Seedata,restrictingandsortingrules
forcreatingviews,395
useexternaltables,91
viewasnamed,391,392
selection,defined,583
selectivity,defined,583
self-join
defined,583
FIRSTandLASTfunctionsin,293
overviewof,333–335
semicolonterminationcharacter,SQLvs.SQL*Plus,54–55,180,534
semijoin,363,583
sequence
createanddrop,408–409
defined,44,583
overviewof,407–408
asschemaobject,392
use,409–410
session,defined,583
SESSION_PRIVS,datadictionaryview,487,548
SESSION_ROLES,datadictionaryview,552
SETclause
setoperatorsvs.,119
UPDATEstatementand,117–119,360–362
SETCOLINVISIBLEONstatement,404
SETcommands,ampersandsubstitution,177–178,180–181
setoperators
combinemultiplequeriesintosinglequery,452–455
controlorderofrowsreturned,459–461
defined,583
descriptionof,450–451
ensureSELECTcombineswith,451–452
INTERSECT,456–457
MINUS,457–458
multiplecombinationsof,458–459
overviewof,450
preventsINSERT/UPDATE/DELETEonviews,398
reviewcertificationsummary,462
reviewdrill,463
reviewself-testQ&A,464–472
SETclausevs.,119
UNIONALL,456
SETUNUSED,columns,84–86
SHOWcommand,ampersandsubstitution,177–178
single-columnindexes,416–420
singlequoteescapecharacter,207–208,252,570
single-row(scalar)functions
character.Seecharacterfunctions
date.Seedatefunctions
defined,280
DUALtableand,206
groupfunctionsvs.,281–283
nesting,234–235
numerical,215–217
overviewof,202
reviewcertificationsummary,235–236
reviewdrill,237
reviewself-testQ&A,238–244
rulesfornesting,300–303
inSELECTstatements,205
typesof,202–205
single-rowsubqueries,356,365–369
SKIP,reservedword,583
skipscanning,419
SocialSecurityNumbers(SSNs),avoidusing,420
SOMEcomparisonoperator,multi-rowsubquery,371
sort
data.Seedata,restrictingandsortingoutputatruntime.Seeampersand
substitutionrows.Seerowsretrievedbyquery,sortSOUNDEXcharacter
function,213–214
specialcharacters,namingdatabaseobjects,48
SQLFundamentalsI(1Z0-051exam),7–12
SQLLanguageReferenceManual
aggregatefunctions,293
databaseobjects,42
privileges,537
asreferenceforthisbook,16
reservedwords,48
SQLfunctions,233
SQLstatementsascommands,29
valuesforNUMBERdatatype,58
SQLstatements
mostcommonlyused,26–27
purposeofDDL,28–29
purposeofDML,29–30
typesofinOracleSQL,27
SQL(StructuredQueryLanguage)
buildSELECTstatement,30–32
defined,583
ERD/relationaldatabaseconnection.SeeERD(entity-relationship
diagram)examoverview.Seeexamoverview
overviewof,2–4
purposeofDDL,28–29
purposeofDML,29–30
relationshipbetweendatabaseand,25–27
reviewcertificationsummary,32
reviewdrill,33–34
reviewself-testQ&A,35–39
rowlimitingclause,182–186
SQL*Plus.Seeampersandsubstitution
SSNs(SocialSecurityNumbers),avoidusing,420
standalonequeries,352
standarddeviation
aggregatefunctionssupporting,281
calculatingSTTDEV,230–232
defined,583
standardtime
daylightsavingtimeas1-houroffsetfrom,573
defined,583
timezonesuseuniform,585
STARTWITHinteger,sequences,408
statements,defined,583
STDDEVanalyticalfunction,230–232
stringdatatype,57
strings,defined,583
structure,ofdatadictionaryviews,475–477
StructuredQueryLanguage.SeeSQL(StructuredQueryLanguage)study
materials,exam,13–15
subjectareas,exam,12–13
subqueries
defined,352,584
multitableINSERTsrequire,499–501,505–511
overviewof,352–353
problemssolvedby,354–355
querydatausingcorrelated,357–359
reviewcertificationsummary,373
reviewdrill,374–376
reviewself-testQ&A,377–388
typesof,356–357
update/deleterowsusingcorrelated,359–363
useEXISTSandNOTEXISTSoperators,363
useWITHclause,364–365
writemultiple-row,369–372
writesingle-row,365–369
subqueryfactoring,WITHclause,364
substitutionvariables,7
SUMfunction,224–226
summaryscreen,exam,6
superaggregate,defined,584
SUSTRcharacterfunction,212–213
synonyms
datadictionaryviews,477
defined,584
dynamicperformance,479
GV$prefixinglobaldynamicperformanceviews,477
renamingdatadictionaryobjectsviapublic,475–477
schemaprefixesand,544–545
SYNONYMobject,44
V$prefixindynamicperformanceviews,477
syntax,defined,584
SYSuseraccount,475,584
SYSDATEdatefunction,217
system-assignednames,53,66–67
systemprivileges
WITHADMINOPTION,539–540
ALLPRIVILEGES,540–541
ANYkeyword,535–538
assigntoroles,550–552
defined,531,584
GRANTandREVOKEonPUBLIC,541–542
grantontablesandusers,542–552
grantingwithGRANTstatement,535–538
listofsome,531–533
overviewof,530–531
prerequisites,533–535
PUBLICkeyword,541–542
viewindatadictionary,547–548
vs.objectprivileges,533
SYSTEMuseraccount,530,584
T
TABLEobject,defined,43
tables
aliases,324–327,330–331,335
constraints,390
createsimple,46–48
datadictionaryinformationstoredin,475
defined,584
external,86–91
indexeson.Seeindexes
inspectindatadictionary,484–485
invisiblecolumnsand,404–405
join.Seejoins
manipulatedata.Seedata,manipulatingmergerowsinto,514–517
naming,48–53
privilegesonuserand,542–552
recoverdropped,425–429
asschemaobjects,390
usesubqueriestopopulate,355
tables,create/manage
categorizeobjects,42–46
constraints,63–80
datatypesforcolumns,56–63
dropcolumns,80–84
externaltables,86–91
overviewof,41
reviewcertificationsummary,91–93
reviewdrill,94–96
reviewself-testQ&A,97–104
setcolumnUNUSED,84–86
simpletables,46–54
structure,54–56
tablespaces,535,584
TCL(TransactionControlLanguage),30,584.Seealsocontroltransactions
technicalsupport,forthisbook,567
testlogistics,examregistration,4
testing
indexontablewithinvisibleindex,421–422
objectprivileges,542–544
text
datatypes,57
defined,584
searchingwithUPPER/LOWER,206–207
third-generationlanguage(3GL),569
thirdnormalform(3NF),23–25
timezone,585
timezonename,585
timezoneoffset,585
TIMESTAMPdatatype
conversionfunctions,260,433–434
dateandNUMBERconstants,223
defined,60
recoverdroppedtablesidentifiedby,425
timestamp,defined,585
TIMESTAMPWITHLOCALTIMEZONEdatatype,61,263
TIMESTAMPWITHTIMEZONEdatatype,60–61,80,261
TIMESTAMP_TO_SCNconversionfunction,433
TOkeyword
FLASHBACKTABLEstatement,426
GRANTstatement,535–538
TO_CHARfunctions
TO_CHAR—CHARACTER,253–254
TO_CHAR—DATE,254–255
TO_CHAR—NUMBER,254
overviewof,253
ROUNDfunctionusedwith,218–219
TO_DATEfunction,259–260
TO_DSINTERVALfunction,262
TO_NUMBERfunction,250–253
top-levelquery,subqueries,352,365
TotalTesterexamsoftware,566
TO_TIMESTAMPnfunction,260,432
TO_TIMESTAMP_TZfunction,261
TO_YMINTERVALfunction,261
TransactionControlLanguage(TCL),30,584.Seealsocontroltransactions
transactions,defined,585
TRIMcharacterfunction,211
TRUNCfunction,216,219–220
TRUNCATEstatement,28,585
TRUNCATETABLEstatement,106–109
U
unconditional,defined,585
unconditionalmultitableINSERTstatements,499,502–505
UNDEFINE,ampersandsubstitution,177
underscore(_)wildcard,LIKEoperator,162–164
undofunction.SeeROLLBACKstatement
undosegments,425,585
Unicode,62,585
UNIONALLoperator,451–452,456,458–459
UNIONoperator
combinemultiplequeriesintosinglequery,452–455
controlorderofrowsreturned,459–461
ensureSELECTcombineswith,451–452
inmultiplecombinations,458–459
assetoperator,450–451
UNIONALLoperatorvs.,456
UNIQUEconstraint
composite,71
createdattimeoftablecreation,71
datatyperestrictionson,79–80
implicitindexcreation,414–415
inPRIMARYKEYconstraint,71
reviewof,79
uniqueindexvs.,420
unique,defined,585
uniqueidentifier,defined,585
uniqueindexes,420
uniquenames,andnamespaces,51–53
UNUSED,setcolumnsto,84–87
UPDATEstatement
COMMIT,multiplesessions,and,125–127
defined,585
asDMLstatement,29
executeonviews,396–400
indexmaintainedwithevery.Seeindexesquerydataviacorrelated
subqueriesin,355,357–359
insingleMERGEstatement,514–516
updaterowsintable,117–121
updaterowswithcorrelatedsubquery,360–362
usesubqueriesin,352–353,355
UPPERcharacterfunction,textsearch,206–207
uppercase,defined,585
useraccesscontrol
privileges,530
privilegesontablesandusers,542–552
privilegesvs.roles,552–553
PUBLICuseraccount,541
reviewcertificationsummary,554
reviewdrill,555–556
reviewself-testQ&A,557–564
systemprivileges.Seesystemprivilegesuseraccount
defined,585
identifyownedobjectsindatadictionary,482–484
overviewof,45
privilegesontables,542–552
roles,551
schemaobjectsvs.nonschemaobjects,45–46
viewsystemprivileges,548
user-definedfunctions,203
user-definedtypes,61
USERfunction,205
USERobject,44,46
USER_prefix,datadictionaryviews,476–477,478
USER_CATALOGview,datadictionary,482–484
USER_CONSTRAINTS,datadictionaryview,487–488
username,createuseraccount,533–534
USER_OBJECTS,datadictionaryview,484,486
USER_RECYCLEBIN,DROPTABLE,426
USER_SYS_PRIVS,datadictionaryview,486,548
USER_TAB_COLUMNS,datadictionaryview,484–485
USER_TABLESview,datadictionary,475–476,484–485
USER_TAB_PRIVS,datadictionaryview,486,548
USER_VIEW,datadictionaryview,486
USINGkeyword,329–330,514–516
UTC(CoordinatedUniversalTime),583,585
utilities,Oracle,88,124
V
V$prefix,dynamicperformancesynonyms,478–479
V_$prefix,dynamicperformanceviews,477–478
VALUESclause,multitableINSERTs,499–500,503–505
VARCHAR2datatype,57
variable,defined,585
variance
aggregratefunctionsfor,305
calculatestandarddeviationand,230–232
defined,585
VARIANCEfunction,withSTDDEV,230–232
VERIFY,ampersandsubstitution,178
views.Seealsodatadictionaryviews
ALTERVIEWstatement,402–403
andconstraints,397
create,393–397
create/dropsequences,408–410
defined,44,585
FIRSTandLASTfunctionsin,293
inline,400–402
INSERTstatementwith,109
ofmultipletables.Seejoins
overviewof,392–393
recompileinvalid,402–403
reviewcertificationsummary,435
reviewdrill,437
reviewself-testQ&A,440–442,446–447
asschemaobjects,391
UPDATEstatementwith,117
updateable,397–400
usingsubqueries,353
visible/invisiblecolumnsimpacting,403–407
visiblecolumns,403–407
visibleindexes,421–422
VISIBLEkeyword,CREATEINDEX,422
W
WHENclause
conditionalmultitableINSERTwith,499–502,505–511
unconditionalmultitableINSERTomits,502–505
WHEN/THENcomparisonpair,CASE,265
WHEREclause
additionalfeaturesof,169–172
benefitsofindexes,413
compareexpressionswith,159–160
customize,172
datatypecomparisonsand,160–162
DELETEstatementwith,121–122,362–363
innerjoinswith,324
LIKEcomparisonoperatorwith,162–164
limitrowsretrievedbyquerywith,158–159
mergerowsintotablewith,514–516
supportforBooleanlogic,164–169
UPDATEstatementwith,117–118,120–121,360–362
usemultiplesubquerieswithin,368
useSQLrowlimitingclauseafter,182–186
usesubqueries,353
wildcardsearches,162–164,369
window,defined,585
wintertimezone.Seestandardtime
WITHADMINOPTIONclause,GRANT,539–540,551
WITHclause,364–365
WITHGRANTOPTIONclause,GRANT,545–546
WITHTIES,rowlimiting,183–186

Navigation menu