大學(xué)畢業(yè)設(shè)計(jì)倉庫管理系統(tǒng)數(shù)據(jù)庫計(jì)算機(jī)外文參考文獻(xiàn)原文及翻譯

1、河北工程大學(xué)畢業(yè)論文（設(shè)計(jì)）論文題目：鴻海種業(yè)倉庫管理系統(tǒng)的 設(shè)計(jì)與實(shí)現(xiàn) 作者姓名： 石成華 專業(yè)班級(jí)： 信管1001 學(xué)號(hào)信息： 100340119 指導(dǎo)老師： 張貴煒 論文日期： 2014.04.10 英文參考文獻(xiàn)原文復(fù)印件及譯文數(shù)據(jù)倉庫數(shù)據(jù)倉庫為商務(wù)運(yùn)作提供結(jié)構(gòu)與工具，以便系統(tǒng)地組織、理解和使用數(shù)據(jù)進(jìn)行決策。大量組織機(jī)構(gòu)已經(jīng)發(fā)現(xiàn)，在當(dāng)今這個(gè)充滿競(jìng)爭(zhēng)、快速發(fā)展的世界，數(shù)據(jù)倉庫是一個(gè)有價(jià)值的工具。在過去的幾年中，許多公司已花費(fèi)數(shù)百萬美元，建立企業(yè)范圍的數(shù)據(jù)倉庫。許多人感到，隨著工業(yè)競(jìng)爭(zhēng)的加劇，數(shù)據(jù)倉庫成了必備的最新營(yíng)銷武器通過更多地了解客戶需求而保住客戶的途徑。“那么”，你可能會(huì)充滿神秘地問

2、，“到底什么是數(shù)據(jù)倉庫？”數(shù)據(jù)倉庫已被多種方式定義，使得很難嚴(yán)格地定義它。寬松地講，數(shù)據(jù)倉庫是一個(gè)數(shù)據(jù)庫，它與組織機(jī)構(gòu)的操作數(shù)據(jù)庫分別維護(hù)。數(shù)據(jù)倉庫系統(tǒng)允許將各種應(yīng)用系統(tǒng)集成在一起，為統(tǒng)一的歷史數(shù)據(jù)分析提供堅(jiān)實(shí)的平臺(tái)，對(duì)信息處理提供支持。按照w.h.inmon，一位數(shù)據(jù)倉庫系統(tǒng)構(gòu)造方面的領(lǐng)頭建筑師的說法，“數(shù)據(jù)倉庫是一個(gè)面向主題的、集成的、時(shí)變的、非易失的數(shù)據(jù)集合，支持管理決策制定”。這個(gè)簡(jiǎn)短、全面的定義指出了數(shù)據(jù)倉庫的主要特征。四個(gè)關(guān)鍵詞，面向主題的、集成的、時(shí)變的、非易失的，將數(shù)據(jù)倉庫與其它數(shù)據(jù)存儲(chǔ)系統(tǒng)（如，關(guān)系數(shù)據(jù)庫系統(tǒng)、事務(wù)處理系統(tǒng)、和文件系統(tǒng)）相區(qū)別。讓我們進(jìn)一步看看這些關(guān)鍵特征。(

3、1) 面向主題的：數(shù)據(jù)倉庫圍繞一些主題，如顧客、供應(yīng)商、產(chǎn)品和銷售組織。數(shù)據(jù)倉庫關(guān)注決策者的數(shù)據(jù)建模與分析，而不是構(gòu)造組織機(jī)構(gòu)的日常操作和事務(wù)處理。因此，數(shù)據(jù)倉庫排除對(duì)于決策無用的數(shù)據(jù)，提供特定主題的簡(jiǎn)明視圖。(2) 集成的：通常，構(gòu)造數(shù)據(jù)倉庫是將多個(gè)異種數(shù)據(jù)源，如關(guān)系數(shù)據(jù)庫、一般文件和聯(lián)機(jī)事務(wù)處理記錄，集成在一起。使用數(shù)據(jù)清理和數(shù)據(jù)集成技術(shù)，確保命名約定、編碼結(jié)構(gòu)、屬性度量的一致性等。(3) 時(shí)變的：數(shù)據(jù)存儲(chǔ)從歷史的角度（例如，過去5-10年）提供信息。數(shù)據(jù)倉庫中的關(guān)鍵結(jié)構(gòu)，隱式或顯式地包含時(shí)間元素。(4) 非易失的：數(shù)據(jù)倉庫總是物理地分離存放數(shù)據(jù)；這些數(shù)據(jù)源于操作環(huán)境下的應(yīng)用數(shù)據(jù)。由于這種

4、分離，數(shù)據(jù)倉庫不需要事務(wù)處理、恢復(fù)和并行控制機(jī)制。通常，它只需要兩種數(shù)據(jù)訪問：數(shù)據(jù)的初始化裝入和數(shù)據(jù)訪問。概言之，數(shù)據(jù)倉庫是一種語義上一致的數(shù)據(jù)存儲(chǔ)，它充當(dāng)決策支持?jǐn)?shù)據(jù)模型的物理實(shí)現(xiàn)，并存放企業(yè)決策所需信息。數(shù)據(jù)倉庫也常常被看作一種體系結(jié)構(gòu)，通過將異種數(shù)據(jù)源中的數(shù)據(jù)集成在一起而構(gòu)造，支持結(jié)構(gòu)化和啟發(fā)式查詢、分析報(bào)告和決策制定?！昂谩保悻F(xiàn)在問，“那么，什么是建立數(shù)據(jù)倉庫？”根據(jù)上面的討論，我們把建立數(shù)據(jù)倉庫看作構(gòu)造和使用數(shù)據(jù)倉庫的過程。數(shù)據(jù)倉庫的構(gòu)造需要數(shù)據(jù)集成、數(shù)據(jù)清理、和數(shù)據(jù)統(tǒng)一。利用數(shù)據(jù)倉庫常常需要一些決策支持技術(shù)。這使得“知識(shí)工人”（例如，經(jīng)理、分析人員和主管）能夠使用數(shù)據(jù)倉庫，快捷、

5、方便地得到數(shù)據(jù)的總體視圖，根據(jù)數(shù)據(jù)倉庫中的信息做出準(zhǔn)確的決策。有些作者使用術(shù)語“建立數(shù)據(jù)倉庫”表示構(gòu)造數(shù)據(jù)倉庫的過程，而用術(shù)語“倉庫dbms”表示管理和使用數(shù)據(jù)倉庫。我們將不區(qū)分二者。 “組織機(jī)構(gòu)如何使用數(shù)據(jù)倉庫中的信息？”許多組織機(jī)構(gòu)正在使用這些信息支持商務(wù)決策活動(dòng)，包括:(1)、增加顧客關(guān)注，包括分析顧客購買模式（如，喜愛買什么、購買時(shí)間、預(yù)算周期、消費(fèi)習(xí)慣）；(2)、根據(jù)季度、年、地區(qū)的營(yíng)銷情況比較，重新配置產(chǎn)品和管理投資，調(diào)整生產(chǎn)策略；(3)、分析運(yùn)作和查找利潤(rùn)源； (4)、管理顧客關(guān)系、進(jìn)行環(huán)境調(diào)整、管理合股人的資產(chǎn)開銷。從異種數(shù)據(jù)庫集成的角度看，數(shù)據(jù)倉庫也是十分有用的。許多組織收集

6、了形形色色數(shù)據(jù)，并由多個(gè)異種的、自治的、分布的數(shù)據(jù)源維護(hù)大型數(shù)據(jù)庫。集成這些數(shù)據(jù)，并提供簡(jiǎn)便、有效的訪問是非常希望的，并且也是一種挑戰(zhàn)。數(shù)據(jù)庫工業(yè)界和研究界都正朝著實(shí)現(xiàn)這一目標(biāo)竭盡全力。 對(duì)于異種數(shù)據(jù)庫的集成，傳統(tǒng)的數(shù)據(jù)庫做法是：在多個(gè)異種數(shù)據(jù)庫上，建立一個(gè)包裝程序和一個(gè)集成程序（或仲裁程序）。這方面的例子包括ibm 的數(shù)據(jù)連接程序和informix的數(shù)據(jù)刀。當(dāng)一個(gè)查詢提交客戶站點(diǎn)，首先使用元數(shù)據(jù)字典對(duì)查詢進(jìn)行轉(zhuǎn)換，將它轉(zhuǎn)換成相應(yīng)異種站點(diǎn)上的查詢。然后，將這些查詢映射和發(fā)送到局部查詢處理器。由不同站點(diǎn)返回的結(jié)果被集成為全局回答。這種查詢驅(qū)動(dòng)的方法需要復(fù)雜的信息過濾和集成處理，并且與局部數(shù)據(jù)源上

7、的處理競(jìng)爭(zhēng)資源。這種方法是低效的，并且對(duì)于頻繁的查詢，特別是需要聚集操作的查詢，開銷很大。 對(duì)于異種數(shù)據(jù)庫集成的傳統(tǒng)方法，數(shù)據(jù)倉庫提供了一個(gè)有趣的替代方案。數(shù)據(jù)倉庫使用更新驅(qū)動(dòng)的方法，而不是查詢驅(qū)動(dòng)的方法。這種方法將來自多個(gè)異種源的信息預(yù)先集成，并存儲(chǔ)在數(shù)據(jù)倉庫中，供直接查詢和分析。與聯(lián)機(jī)事務(wù)處理數(shù)據(jù)庫不同，數(shù)據(jù)倉庫不包含最近的信息。然而，數(shù)據(jù)倉庫為集成的異種數(shù)據(jù)庫系統(tǒng)帶來了高性能，因?yàn)閿?shù)據(jù)被拷貝、預(yù)處理、集成、注釋、匯總，并重新組織到一個(gè)語義一致的數(shù)據(jù)存儲(chǔ)中。在數(shù)據(jù)倉庫中進(jìn)行的查詢處理并不影響在局部源上進(jìn)行的處理。此外，數(shù)據(jù)倉庫存儲(chǔ)并集成歷史信息，支持復(fù)雜的多維查詢。這樣，建立數(shù)據(jù)倉庫在工業(yè)

8、界已非常流行。 1.操作數(shù)據(jù)庫系統(tǒng)與數(shù)據(jù)倉庫的區(qū)別由于大多數(shù)人都熟悉商品關(guān)系數(shù)據(jù)庫系統(tǒng)，將數(shù)據(jù)倉庫與之比較，就容易理解什么是數(shù)據(jù)倉庫。 聯(lián)機(jī)操作數(shù)據(jù)庫系統(tǒng)的主要任務(wù)是執(zhí)行聯(lián)機(jī)事務(wù)和查詢處理。這種系統(tǒng)稱為聯(lián)機(jī)事務(wù)處理（oltp）系統(tǒng)。它們涵蓋了一個(gè)組織的大部分日常操作，如購買、庫存、制造、銀行、工資、注冊(cè)、記帳等。另一方面，數(shù)據(jù)倉庫系統(tǒng)在數(shù)據(jù)分析和決策方面為用戶或“知識(shí)工人”提供服務(wù)。這種系統(tǒng)可以用不同的格式組織和提供數(shù)據(jù)，以便滿足不同用戶的形形色色需求。這種系統(tǒng)稱為聯(lián)機(jī)分析處理（olap）系統(tǒng)。 oltp 和olap 的主要區(qū)別概述如下。 (1) 用戶和系統(tǒng)的面向性：oltp 是面向顧客的，用

9、于辦事員、客戶、和信息技術(shù)專業(yè)人員的事務(wù)和查詢處理。olap 是面向市場(chǎng)的，用于知識(shí)工人（包括經(jīng)理、主管、和分析人員）的數(shù)據(jù)分析。 (2) 數(shù)據(jù)內(nèi)容：oltp 系統(tǒng)管理當(dāng)前數(shù)據(jù)。通常，這種數(shù)據(jù)太瑣碎，難以方便地用于決策。olap 系統(tǒng)管理大量歷史數(shù)據(jù)，提供匯總和聚集機(jī)制，并在不同的粒度級(jí)別上存儲(chǔ)和管理信息。這些特點(diǎn)使得數(shù)據(jù)容易用于見多識(shí)廣的決策。 (3) 數(shù)據(jù)庫設(shè)計(jì)：通常，oltp 系統(tǒng)采用實(shí)體-聯(lián)系（er）模型和面向應(yīng)用的數(shù)據(jù)庫設(shè)計(jì)。而olap 系統(tǒng)通常采用星形或雪花模型和面向主題的數(shù)據(jù)庫設(shè)計(jì)。 (4) 視圖：oltp 系統(tǒng)主要關(guān)注一個(gè)企業(yè)或部門內(nèi)部的當(dāng)前數(shù)據(jù)，而不涉及歷史數(shù)據(jù)或不同組織的數(shù)

10、據(jù)。相比之下，由于組織的變化，olap 系統(tǒng)常?？缭綌?shù)據(jù)庫模式的多個(gè)版本。olap 系統(tǒng)也處理來自不同組織的信息，由多個(gè)數(shù)據(jù)存儲(chǔ)集成的信息。由于數(shù)據(jù)量巨大，olap 數(shù)據(jù)也存放在多個(gè)存儲(chǔ)介質(zhì)上。 (5)、訪問模式：oltp 系統(tǒng)的訪問主要由短的、原子事務(wù)組成。這種系統(tǒng)需要并行控制和恢復(fù)機(jī)制。然而，對(duì)olap系統(tǒng)的訪問大部分是只讀操作（由于大部分?jǐn)?shù)據(jù)倉庫存放歷史數(shù)據(jù)，而不是當(dāng)前數(shù)據(jù)），盡管許多可能是復(fù)雜的查詢。 oltp 和olap 的其它區(qū)別包括數(shù)據(jù)庫大小、操作的頻繁程度、性能度量等。2.但是，為什么需要一個(gè)分離的數(shù)據(jù)倉庫“既然操作數(shù)據(jù)庫存放了大量數(shù)據(jù)”，你注意到，“為什么不直接在這種數(shù)據(jù)庫上

11、進(jìn)行聯(lián)機(jī)分析處理，而是另外花費(fèi)時(shí)間和資源去構(gòu)造一個(gè)分離的數(shù)據(jù)倉庫？”分離的主要原因是提高兩個(gè)系統(tǒng)的性能。操作數(shù)據(jù)庫是為已知的任務(wù)和負(fù)載設(shè)計(jì)的，如使用主關(guān)鍵字索引和散列，檢索特定的記錄，和優(yōu)化“罐裝的”查詢。另一方面，數(shù)據(jù)倉庫的查詢通常是復(fù)雜的，涉及大量數(shù)據(jù)在匯總級(jí)的計(jì)算，可能需要特殊的數(shù)據(jù)組織、存取方法和基于多維視圖的實(shí)現(xiàn)方法。在操作數(shù)據(jù)庫上處理olap查詢，可能會(huì)大大降低操作任務(wù)的性能。此外，操作數(shù)據(jù)庫支持多事務(wù)的并行處理，需要加鎖和日志等并行控制和恢復(fù)機(jī)制，以確保一致性和事務(wù)的強(qiáng)健性。通常，olap查詢只需要對(duì)數(shù)據(jù)記錄進(jìn)行只讀訪問，以進(jìn)行匯總和聚集。如果將并行控制和恢復(fù)機(jī)制用于這olap操

12、作，就會(huì)危害并行事務(wù)的運(yùn)行，從而大大降低oltp系統(tǒng)的吞吐量。最后，數(shù)據(jù)倉庫與操作數(shù)據(jù)庫分離是由于這兩種系統(tǒng)中數(shù)據(jù)的結(jié)構(gòu)、內(nèi)容和用法都不相同。決策支持需要?dú)v史數(shù)據(jù)，而操作數(shù)據(jù)庫一般不維護(hù)歷史數(shù)據(jù)。在這種情況下，操作數(shù)據(jù)庫中的數(shù)據(jù)盡管很豐富，但對(duì)于決策，常常還是遠(yuǎn)遠(yuǎn)不夠的。決策支持需要將來自異種源的數(shù)據(jù)統(tǒng)一（如，聚集和匯總），產(chǎn)生高質(zhì)量的、純凈的和集成的數(shù)據(jù)。相比之下，操作數(shù)據(jù)庫只維護(hù)詳細(xì)的原始數(shù)據(jù)（如事務(wù)），這些數(shù)據(jù)在進(jìn)行分析之前需要統(tǒng)一。由于兩個(gè)系統(tǒng)提供很不相同的功能，需要不同類型的數(shù)據(jù)，因此需要維護(hù)分離的數(shù)據(jù)庫。datawarehousingprovidesarchitecturesand

13、toolsforbusinessexecutivestosystematicallyorganize,understand,andusetheirdatatomakestrategicdecisions.alargenumberoforganizationshavefoundthatdatawarehousesystemsarevaluabletoolsintodayscompetitive,fastevolvingworld.inthelastseveralyears,manyfirmshavespentmillionsofdollarsinbuildingenterprise-wideda

14、tawarehouses.manypeoplefeelthatwithcompetitionmountingineveryindustry,datawarehousingisthelatestmust-havemarketingweaponawaytokeepcustomersbylearningmoreabouttheirneeds.“so,youmayask,fullofintrigue,“whatexactlyisadatawarehouse?datawarehouseshavebeendefinedinmanyways,makingitdifficulttoformulatearigo

15、rousdefinition.looselyspeaking,adatawarehousereferstoadatabasethatismaintainedseparatelyfromanorganizationsoperationaldatabases.datawarehousesystemsallowfortheintegrationofavarietyofapplicationsystems.theysupportinformationprocessingbyprovidingasolidplatformofconsolidated,historicaldataforanalysis.a

16、ccordingtow.h.inmon,aleadingarchitectintheconstructionofdatawarehousesystems,“adatawarehouseisasubject-oriented,integrated,time-variant,andnonvolatilecollectionofdatainsupportofmanagementsdecisionmakingprocess.thisshort,butcomprehensivedefinitionpresentsthemajorfeaturesofadatawarehouse.thefourkeywor

17、ds,subject-oriented,integrated,time-variant,andnonvolatile,distinguishdatawarehousesfromotherdatarepositorysystems,suchasrelationaldatabasesystems,transactionprocessingsystems,andfilesystems.letstakeacloserlookateachofthesekeyfeatures.(1).subject-oriented:adatawarehouseisorganizedaroundmajorsubjects

18、,suchascustomer,vendor,product,andsales.ratherthanconcentratingontheday-to-dayoperationsandtransactionprocessingofanorganization,adatawarehousefocusesonthemodelingandanalysisofdatafordecisionmakers.hence,datawarehousestypicallyprovideasimpleandconciseviewaroundparticularsubjectissuesbyexcludingdatat

19、hatarenotusefulinthedecisionsupportprocess.(2)integrated:adatawarehouseisusuallyconstructedbyintegratingmultipleheterogeneoussources,suchasrelationaldatabases,flatfiles,andon-linetransactionrecords.datacleaninganddataintegrationtechniquesareappliedtoensureconsistencyinnamingconventions,encodingstruc

20、tures,attributemeasures,andsoon.(3).time-variant:dataarestoredtoprovideinformationfromahistoricalperspective(e.g.,thepast5-10years).everykeystructureinthedatawarehousecontains,eitherimplicitlyorexplicitly,anelementoftime.(4)nonvolatile:adatawarehouseisalwaysaphysicallyseparatestoreofdatatransformedf

21、romtheapplicationdatafoundintheoperationalenvironment.duetothisseparation,adatawarehousedoesnotrequiretransactionprocessing,recovery,andconcurrencycontrolmechanisms.itusuallyrequiresonlytwooperationsindataaccessing:initialloadingofdataandaccessofdata.insum,adatawarehouseisasemanticallyconsistentdata

22、storethatservesasaphysicalimplementationofadecisionsupportdatamodelandstorestheinformationonwhichanenterpriseneedstomakestrategicdecisions.adatawarehouseisalsooftenviewedasanarchitecture,constructedbyintegratingdatafrommultipleheterogeneoussourcestosupportstructuredand/oradhocqueries,analyticalrepor

23、ting,anddecisionmaking.“ok,younowask,“what,then,isdatawarehousing?basedontheabove,weviewdatawarehousingastheprocessofconstructingandusingdatawarehouses.theconstructionofadatawarehouserequiresdataintegration,datacleaning,anddataconsolidation.theutilizationofadatawarehouseoftennecessitatesacollectiono

24、fdecisionsupporttechnologies.thisallows“knowledgeworkers(e.g.,managers,analysts,andexecutives)tousethewarehousetoquicklyandconvenientlyobtainanoverviewofthedata,andtomakesounddecisionsbasedoninformationinthewarehouse.someauthorsusetheterm“datawarehousingtoreferonlytotheprocessofdatawarehouseconstruc

25、tion,whilethetermwarehousedbmsisusedtorefertothemanagementandutilizationofdatawarehouses.wewillnotmakethisdistinctionhere.“howareorganizationsusingtheinformationfromdatawarehouses?manyorganizationsareusingthisinformationtosupportbusinessdecisionmakingactivities,including:(1)increasingcustomerfocus,w

26、hichincludestheanalysisofcustomerbuyingpatterns(suchasbuyingpreference,buyingtime,budgetcycles,andappetitesforspending),(2)repositioningproductsandmanagingproductportfoliosbycomparingtheperformanceofsalesbyquarter,byyear,andbygeographicregions,inordertofine-tuneproductionstrategies,(3)analyzingopera

27、tionsandlookingforsourcesofprofit,(4)managingthecustomerrelationships,makingenvironmentalcorrections,andmanagingthecostofcorporateassets.datawarehousingisalsoveryusefulfromthepointofviewofheterogeneousdatabaseintegration.manyorganizationstypicallycollectdiversekindsofdataandmaintainlargedatabasesfro

28、mmultiple,heterogeneous,autonomous,anddistributedinformationsources.tointegratesuchdata,andprovideeasyandefficientaccesstoitishighlydesirable,yetchallenging.muchefforthasbeenspentinthedatabaseindustryandresearchcommunitytowardsachievingthisgoal.thetraditionaldatabaseapproachtoheterogeneousdatabasein

29、tegrationistobuildwrappersandintegrators(ormediators)ontopofmultiple,heterogeneousdatabases.avarietyofdatajoineranddatabladeproductsbelongtothiscategory.whenaqueryisposedtoaclientsite,ametadatadictionaryisusedtotranslatethequeryintoqueriesappropriatefortheindividualheterogeneoussitesinvolved.thesequ

30、eriesarethenmappedandsenttolocalqueryprocessors.theresultsreturnedfromthedifferentsitesareintegratedintoaglobalanswerset.thisquery-drivenapproachrequirescomplexinformationfilteringandintegrationprocesses,andcompetesforresourceswithprocessingatlocalsources.itisinefficientandpotentiallyexpensiveforfre

31、quentqueries,especiallyforqueriesrequiringaggregations.datawarehousingprovidesaninterestingalternativetothetraditionalapproachofheterogeneousdatabaseintegrationdescribedabove.ratherthanusingaquery-drivenapproach,datawarehousingemploysanupdate-drivenapproachinwhichinformationfrommultiple,heterogeneou

32、ssourcesisintegratedinadvanceandstoredinawarehousefordirectqueryingandanalysis.unlikeon-linetransactionprocessingdatabases,datawarehousesdonotcontainthemostcurrentinformation.however,adatawarehousebringshighperformancetotheintegratedheterogeneousdatabasesystemsincedataarecopied,preprocessed,integrat

33、ed,annotated,summarized,andrestructuredintoonesemanticdatastore.furthermore,queryprocessingindatawarehousesdoesnotinterferewiththeprocessingatlocalsources.moreover,datawarehousescanstoreandintegratehistoricalinformationandsupportcomplexmultidimensionalqueries.asaresult,datawarehousinghasbecomeverypo

34、pularinindustry.1.differencesbetweenoperationaldatabasesystemsanddatawarehousessincemostpeoplearefamiliarwithcommercialrelationaldatabasesystems,itiseasytounderstandwhatadatawarehouseisbycomparingthesetwokindsofsystems.themajortaskofon-lineoperationaldatabasesystemsistoperformon-linetransactionandqu

35、eryprocessing.thesesystemsarecalledon-linetransactionprocessing(oltp)systems.theycovermostoftheday-to-dayoperationsofanorganization,suchas,purchasing,inventory,manufacturing,banking,payroll,registration,andaccounting.datawarehousesystems,ontheotherhand,serveusersor“knowledgeworkersintheroleofdataana

36、lysisanddecisionmaking.suchsystemscanorganizeandpresentdatainvariousformatsinordertoaccommodatethediverseneedsofthedifferentusers.thesesystemsareknownason-lineanalyticalprocessing(olap)systems.themajordistinguishingfeaturesbetweenoltpandolaparesummarizedasfollows.(1).usersandsystemorientation:anoltp

37、systemiscustomer-orientedandisusedfortransactionandqueryprocessingbyclerks,clients,andinformationtechnologyprofessionals.anolapsystemismarket-orientedandisusedfordataanalysisbyknowledgeworkers,includingmanagers,executives,andanalysts.(2).datacontents:anoltpsystemmanagescurrentdatathat,typically,aret

38、oodetailedtobeeasilyusedfordecisionmaking.anolapsystemmanageslargeamountsofhistoricaldata,providesfacilitiesforsummarizationandaggregation,andstoresandmanagesinformationatdifferentlevelsofgranularity.thesefeaturesmakethedataeasierforuseininformeddecisionmaking.(3).databasedesign:anoltpsystemusuallya

39、doptsanentity-relationship(er)datamodelandanapplication-orienteddatabasedesign.anolapsystemtypicallyadoptseitherastarorsnowflakemodel,andasubject-orienteddatabasedesign.(4).view:anoltpsystemfocusesmainlyonthecurrentdatawithinanenterpriseordepartment,withoutreferringtohistoricaldataordataindifferento

40、rganizations.incontrast,anolapsystemoftenspansmultipleversionsofadatabaseschema,duetotheevolutionaryprocessofanorganization.olapsystemsalsodealwithinformationthatoriginatesfromdifferentorganizations,integratinginformationfrommanydatastores.becauseoftheirhugevolume,olapdataarestoredonmultiplestoragem

41、edia.(5).accesspatterns:theaccesspatternsofanoltpsystemconsistmainlyofshort,atomictransactions.suchasystemrequiresconcurrencycontrolandrecoverymechanisms.however,accessestoolapsystemsaremostlyread-onlyoperations(sincemostdatawarehousesstorehistoricalratherthanup-to-dateinformation),althoughmanycould

42、becomplexqueries.otherfeatureswhichdistinguishbetweenoltpandolapsystemsincludedatabasesize,frequencyofoperations,andperformancemetricsandsoon.2.but,whyhaveaseparatedatawarehouse?“sinceoperationaldatabasesstorehugeamountsofdata,youobserve,“whynotperformon-lineanalyticalprocessingdirectlyonsuchdatabas

43、esinsteadofspendingadditionaltimeandresourcestoconstructaseparatedatawarehouse?amajorreasonforsuchaseparationistohelppromotethehighperformanceofbothsystems.anoperationaldatabaseisdesignedandtunedfromknowntasksandworkloads,suchasindexingandhashingusingprimarykeys,searchingforparticularrecords,andopti

44、mizing“cannedqueries.ontheotherhand,datawarehousequeriesareoftencomplex.theyinvolvethecomputationoflargegroupsofdataatsummarizedlevels,andmayrequiretheuseofspecialdataorganization,access,andimplementationmethodsbasedonmultidimensionalviews.processingolapqueriesinoperationaldatabaseswouldsubstantiall

45、ydegradetheperformanceofoperationaltasks.moreover,anoperationaldatabasesupportstheconcurrentprocessingofseveraltransactions.concurrencycontrolandrecoverymechanisms,suchaslockingandlogging,arerequiredtoensuretheconsistencyandrobustnessoftransactions.anolapqueryoftenneedsread-onlyaccessofdatarecordsfo

46、rsummarizationandaggregation.concurrencycontrolandrecoverymechanisms,ifappliedforsucholapoperations,mayjeopardizetheexecutionofconcurrenttransactionsandthussubstantiallyreducethethroughputofanoltpsystem.finally,theseparationofoperationaldatabasesfromdatawarehousesisbasedonthedifferentstructures,cont

47、ents,andusesofthedatainthesetwosystems.decisionsupportrequireshistoricaldata,whereasoperationaldatabasesdonottypicallymaintainhistoricaldata.inthiscontext,thedatainoperationaldatabases,thoughabundant,isusuallyfarfromcompletefordecisionmaking.decisionsupportrequiresconsolidation(suchasaggregationands

48、ummarization)ofdatafromheterogeneoussources,resultinginhighquality,cleansedandintegrateddata.incontrast,operationaldatabasescontainonlydetailedrawdata,suchastransactions,whichneedtobeconsolidatedbeforeanalysis.sincethetwosystemsprovidequitedifferentfunctionalitiesandrequiredifferentkindsofdata,itisn

49、ecessarytomaintainseparatedatabases.薃肀莂蒃袂肀肂蠆袈聿芄薂螄肈莇螇蝕肇葿薀罿肆腿莃裊肅芁薈螁膄莃莁蚇膄肅薇薃膃芅荿羈膂莈蚅袇膁蒀蒈螃膀膀蚃蠆腿節(jié)蒆羈羋莄蟻襖羋蒆蒄螀芇膆蝕蚆袃莈蒃螞袂蒁螈羀袁膀薁袆袁芃螆螂袀蒞蕿蚈衿蕆莂羇羈膇薇袃羇艿莀蝿羆蒂薆螅羅膁蒈蟻羅芄蚄罿羄莆蕆裊羃蒈螞螁羂膈蒅蚇肁芀蟻薃肀莂蒃袂肀肂蠆袈聿芄薂螄肈莇螇蝕肇葿薀罿肆腿莃裊肅芁薈螁膄莃莁蚇膄肅薇薃膃芅荿羈膂莈蚅袇膁蒀蒈螃膀膀蚃蠆腿節(jié)蒆羈羋莄蟻襖羋蒆蒄螀芇膆蝕蚆袃莈蒃螞袂蒁螈羀袁膀薁袆袁芃螆螂袀蒞蕿蚈衿蕆莂羇羈膇薇袃羇艿莀蝿羆蒂薆螅羅膁蒈蟻羅芄蚄罿羄莆蕆裊羃蒈螞螁羂膈蒅蚇肁芀蟻薃肀莂蒃袂肀肂蠆

50、袈聿芄薂螄肈莇螇蝕肇葿薀罿肆腿莃裊肅芁薈螁膄莃莁蚇膄肅薇薃膃芅荿螀羀膆蒃蚆肀羋芆薂聿羈蒂蒈肈肀芅袆?wù)仄M薀螂肆蒞莃蚈肅肅薈薄螞膇莁蒀蟻艿薇蝿螀罿荿蚅蝿肁薅薁螈膄莈薇螈莆膀袆螇肆蒆螁螆膈艿蚇螅芀蒄薃螄羀芇葿袃肂蒃螈袂膄芅蚄袂芇蒁蝕袁肆芄薆袀腿蕿蒂衿芁莂螁袈羈薇蚇袇肅莀薃羆膅薆葿羆羋荿螇羅羇膁螃羄膀莇蠆羃節(jié)芀薅羂羂蒅蒁羈肄羋螀羀膆蒃蚆肀羋芆薂聿羈蒂蒈肈肀芅袆?wù)仄M薀螂肆蒞莃蚈肅肅薈薄螞膇莁蒀蟻艿薇蝿螀罿荿蚅蝿肁薅薁螈膄莈薇螈莆膀袆螇肆蒆螁螆膈艿蚇螅芀蒄薃螄羀芇葿袃肂蒃螈袂膄芅蚄袂芇蒁蝕袁肆芄薆袀腿蕿蒂衿芁莂螁袈羈薇蚇袇肅莀薃羆膅薆葿羆羋荿螇羅羇膁螃羄膀莇蠆羃節(jié)芀薅羂羂蒅蒁羈肄羋螀羀膆蒃蚆肀羋芆薂聿羈蒂

51、蒈肈肀芅袆?wù)仄M薀螂肆蒞莃蚈肅肅薈薄螞膇莁蒀蟻艿薇蝿螀罿荿蚅蝿肁薅薁螈膄莈薇螈莆膀袆螇肆蒆螁螆膈艿蚇螅芀蒄薃螄羀芇葿袃肂蒃螈袂膄芅蚄袂芇蒁蝕袁肆芄薆袀腿蕿蒂衿芁莂螁袈羈薇蚇袇肅莀薃羆膅薆葿羆羋荿螇羅羇膁螃羄膀莇蠆羃節(jié)芀薅羂羂蒅蒁羈肄羋螀羀膆蒃蚆肀羋芆薂聿羈蒂蒈肈肀芅袆?wù)仄M薀螂肆蒞莃蚈肅肅薈薄螞膇莁蒀蟻艿薇蝿螀罿荿蚅蝿肁薅薁螈膄莈薇螈莆膀袆螇肆蒆螁螆膈艿蚇螅芀蒄薃螄羀芇葿袃肂蒃螈袂膄芅蚄袂芇蒁蝕袁肆芄薆袀腿蕿蒂衿芁莂螁袈羈薇蚇袇肅莀薃羆膅薆葿羆羋荿螇羅羇膁螃羄膀莇蠆羃節(jié)芀薅羂羂蒅蒁羈肄羋螀羀膆蒃蚆肀羋芆薂聿羈蒂蒈肈肀芅袆?wù)仄M薀螂肆蒞莃蚈肅肅薈薄螞膇莁蒀蟻艿薇蝿螀罿荿蚅蝿肁薅薁螈膄莈薇螈莆膀袆螇肆蒆

52、螁螆膈艿蚇螅芀蒄薃螄羀芇葿袃肂蒃螈袂膄芅蚄袂芇蒁蝕袁肆芄薆袀腿蕿蒂衿芁莂螁袈羈薇蚇袇肅莀薃羆膅薆葿羆羋荿螇羅羇膁螃羄膀莇蠆羃節(jié)芀薅羂羂蒅蒁羈肄羋螀羀膆蒃蚆肀羋芆薂聿羈蒂蒈肈肀芅袆?wù)仄M薀螂肆蒞莃蚈肅肅薈薄螞膇莁蒀蟻艿薇蝿螀罿荿蚅蝿肁薅薁螈膄莈薇螈莆膀袆螇肆蒆螁螆膈艿蚇螅芀蒄薃螄羀芇葿袃肂蒃螈袂膄芅蚄袂芇蒁蝕袁肆芄薆袀腿蕿蒂衿芁莂螁袈羈薇蚇袇肅莀薃羆膅薆葿羆羋荿螇羅羇膁螃羄膀莇蠆羃節(jié)芀薅羂羂蒅蒁羈肄羋螀羀膆蒃蚆肀羋芆薂聿羈蒂蒈肈肀芅袆?wù)仄M薀螂肆蒞莃蚈肅肅薈薄螞膇莁蒀蟻艿薇蝿螀罿荿蚅蝿肁薅薁螈膄莈薇螈莆膀袆螇肆蒆螁螆膈艿蚇螅芀蒄薃螄羀芇葿袃肂蒃螈袂膄芅蚄袂芇蒁蝕袁肆芄薆袀腿蕿蒂衿芁莂螁袈羈薇蚇袇肅莀

53、薃羆膅薆葿羆羋荿螇羅羇膁螃羄膀莇蠆羃節(jié)芀薅羂羂蒅蒁羈肄羋螀羀膆蒃蚆肀羋芆薂聿羈蒂蒈肈肀芅袆?wù)仄M薀螂肆蒞莃蚈肅肅薈薄螞膇莁蒀蟻艿薇蝿螀罿荿蚅蝿肁薅薁螈膄莈薇螈莆膀袆螇肆蒆螁螆膈艿蚇螅芀蒄薃螄羀芇葿袃肂蒃螈袂膄芅蚄袂芇蒁蝕袁肆芄薆袀腿蕿蒂衿芁莂螁袈羈薇蚇袇肅莀薃羆膅薆葿羆羋荿螇羅羇膁螃羄膀莇蠆羃節(jié)芀薅羂羂蒅蒁羈肄羋螀羀膆蒃蚆肀羋芆薂聿羈蒂蒈肈肀芅袆?wù)仄M薀螂肆蒞莃蚈肅肅薈薄螞膇莁蒀蟻艿薇蝿螀罿荿蚅蝿肁薅薁螈膄莈薇螈莆膀袆螇肆蒆螁螆膈艿蚇螅芀蒄薃螄羀芇葿袃肂蒃螈袂膄芅蚄袂芇蒁蝕袁肆芄薆袀腿蕿蒂衿芁莂螁袈羈薇蚇袇肅莀薃羆膅薆葿羆羋荿螇羅羇膁螃羄膀莇蠆羃節(jié)芀薅羂羂蒅蒁羈肄羋螀羀膆蒃蚆肀羋芆薂聿羈蒂蒈肈肀芅袆?wù)仄M薀螂肆蒞莃蚈肅肅薈薄螞膇莁蒀蟻艿薇蝿螀罿荿蚅蝿肁薅薁螈膄莈薇螈莆膀袆螇肆蒆螁螆膈艿蚇螅芀蒄薃螄羀芇葿袃肂蒃螈袂膄芅蚄袂芇蒁蝕袁肆芄薆袀腿蕿蒂衿芁莂螁袈羈薇蚇袇肅莀薃羆膅薆葿羆羋荿螇羅羇膁螃羄膀莇蠆羃節(jié)芀薅羂羂蒅蒁羈肄羋螀羀膆蒃蚆肀羋芆薂聿羈蒂蒈肈肀芅袆?wù)仄M薀螂肆蒞莃蚈肅肅薈