数据仓库维度建模技术及其应用研究

数据仓库现今已被各种规模的机构广泛使用，维度建模技术作为展示数据仓库的主导技术，在数据仓库的应用中起着关键作用．研究数据仓库的维度建模技术，提出了对于变化维度的处理方法，根据维度属性与时间的关系，将维度属性划分成稳定维度、渐变维度与快变维度，阐述了跟踪事实表中历史数据的方法，在渐变维度的处理方法中，讨论了自然关键字与代理关键字的不同作用与具体用法，在快变维度的处理方法中讨论了“微型维度”与“预设波段”概念与作用，并结合基于数据仓库技术的公共课程评估系统，给出了系统数据仓库维度模型的设计方案，对渐度维度与快变维度的处理策略，以及应用相关维度属性进行联机分析处理的结果。

Research and Application of Dimensional Modeling Technoloy in Data Warehouse

The technology of data warehouse is being widely used in decision making and data analysis . Di mensionalmodelingis the data organizingformin data warehouse andit is also a key factor in the data warehouse application. Mostapplications in di mensional modeling put their emphases on changeless di mensions and modifythe values of the di mensionalproperties directly when di mensions are changing.Althoughtracingthe changes accurately are not necessary ,ignoringthechanges completely will invalidate the statistics .Inthis paper an approachtothe changing di mension processing proposed .It concentrates on the two kinds of di mensions ,the slowly changing di mensions and the rapidly changing di mensions . Wealso put forward the solutions to deal withthese two kinds of di mensions ,especially explaining howto use the solutions totrace the history data records from the fact table . Two conceptions are in the slowly changing di mension processingdiscussed one is nature key word and the other is agent key word . On the other hand ,in the rapdly changing di mensionprocessing,two mothods are mentioned :one is called minidi mension and the other is called wave band prefabrication.Inorder to combine theory with practice ,the common curriculumteaching evaluation system (CCTES) based on the datawarehouse is developed , which overcomes the deficiency of the current methods of teaching evaluation, such assubjectivity,isolated viewpoint and no comparison with history data . This paper not only gives the architecture of thesystem,but also describes the design of the systemdi mensional modeling which uses the changing-di mension technologiesmentionedin the paper . Finally , it offers the results of the on-line analytical processing with the slowly changingdi mension,the rapdly changing di mension and the unchanging di mension respectively .