improving+materialized+view+selection+under+storage+constraint

《improving+materialized+view+selection+under+storage+constraint》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、AbstractDatawarehousingandon-lineanalyticalprocessing(otAP)haveincreasinglybecomeacenterofthedatabaseindustry．Theabilityandtheinsurancethatthey西VetocompaniesfortheSuccessoftheiractivitiesategreatlyimproved．Animportantissueindatawarehousedevelopmentisth

2、eselectionofasetofviewstomaterializeinordertoaccelerateOnlineanalyticalprocessingqueries，givencertainspaceandmaintenancetimeconstraints．Ontheotherhand，thematerializedviews(MV)selectionproblemisanNP-·complete；thenumberofpossiblesummary··viewsinthemulti·

3、-dimensional(MD)modelincreasesexplosivelywiththeincreasingnumberofdimensionalattributes．Thehighstoragecostandcomputationcostmakeitunfeasibleforanysystemtomaterializeallofthesepossibleviews．Thus，precisingthesummary-viewsformaterializationbecomesanessent

4、ialchallengeinOLAPresearch．Tosolvethisproblemofsummary—viewsexplosionandstoragecost，wesupplyanimprovedgreedyalgorithmforselectingbestviewsformaterializationbasedontheirweights．Givenaspaceconstraintforthematerializedset，thealgorithmpicksaviewwiththemaxi

5、mumBenefitPerUnitSpace(BPUS)tomaterializefromthecandidateviewset．Afteraperiodoftime(preciselyatthemaintenancetime)，andgivinganewqueryfrequencyforeachview,thealgorithmdeterminetheviewtoberemovedfromthematerializedviewset．Logicallyspeakingthoseviewremove

6、dareviewsthatarelessusefultothedatawarehouse．Thefinalstepistofillthespaceleftbytheremovedviewwiththeothernewoneuntiltheoptimalmaterializedviewsetisreached．Duringtheexperimentofthisalgorithm，weobservethattheafteraperiodoftime，somematerializedviewweights

7、(BPUS)becomeverythinwhileothercandidateviewfromthecandidateviewsetarehavingmoreimportant．Thisnewmethodprovidesawaytosolvethisproblemofuselessmaterializedview．Wethinkthisworkwillcontributegreatlyintheimprovingtheefficiencyofmaterializedviewsusageintheda

8、tawarehouseusage．Keywords：DataWarehouses；OLAP；Cube；MaterializedViewSelection；IIIMASTER’STHESiSLISTOFFIGURESFigurelStarschema⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯．⋯．⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯．．3Figure2ASnowflakeSchema⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯．⋯⋯⋯⋯⋯．．4Figure3AT