数据仓库多维计算的研究与实现

《数据仓库多维计算的研究与实现》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、北京工业大学硕士学位论文数据仓库多维计算的研究与实现姓名：陈耀丰申请学位级别：硕士专业：计算机软件与理论指导教师：周小兵20070401AbstractHowtomanage．and峨alargeamountofdamthatenterpriseproducedinthecouf辩ofrunningericctivelyiSanimportantproblemthatMISstafffaces．TheDataWarehousingandOn-LineAnalyticalProcessing(OLAP)providea

2、solutiontosolvethisproblem．DataWarehousetakesUSCofa‘oncenwatedstructurewhichstoresallthedataofthecompanyinauniformdatabase．Generally,thedamofadamwarehouseareorgaIIi∞dinaformofdatacube，andbenefitsforcombiningandinquiringthedataindifferentangle．Toimprovetheinquir

3、ingandanglicizingspeed，thedatacubeisalwayspre-computedbeforetheOn-LineAnal如calProcessing，whichistheprocessofthedatacubematerialization．Amongthepre-computation,theaggregationcomputationtakeagreatpart．Aggregationcomputationisakindofoperationwhichtakesallimportant

4、roleinthedatawarehousesystem,themainideaofwhichispre-computingthedatausingtheaggregationfunctionsuchascount,sum,average；max,nfin,andthenthe他saltCanbestoreinthedatawarehouseaheadoftime．Asaresult,theOn-LineAnalyticalProcessingCanbebetterdoneonthebaseofthepre-comp

5、uteddata．11"definitionofthedatacubeisthepremisesofthedatacubecomputation．Basedonthedatacubemodelwhichisdefinitedinpaper【29]，wehavedoneresearchonthemulti-dimensionalcomputationofthedatacube．ThenWeproposeourownmulti-dimensionalcomputationdefinitionwhichisusedbetw

6、eentheslicesorthedatacellsandalsothealgorithmbasedontheSQL．Atthe姗etime．wehavedesignedthemultidimensionalcomputationdescriptionlanguageandimplementedthealgorithmbyprograminthecomputerenvironment．Furthermore，wediscusstherelationshipbcDN∞llthemulti-dimensionalcomp

7、utationandaggregationcomputation．Tomakeuptheshortcomingsofouralgorithmaboutaggregationcomputation，especiallyontheholisdcaggregation,wegiveitsimplementationintheOLAPreportprocessing．n圮applicationofthedatacubecomputationalgorithmintheDataWarehouseSystemofBeijingC

8、PCommercialTradingCo．Ltd．isalsomentionedinthis肼Ip％infact,itworkswellinthepracticalapplication．KeywordsDataWarehouse；Datacube；multidimensionalcomputation；Materialization；Aggr