Abstract: The ‘non-volatility’ specialty of data warehouse includes two aspects：one means that the fact data are non-volatility,which can be promised by the ‘Time’ dimension.Another aspect means the schema of data warehouse is also non-volatility,and this has to be promised by schema versioning.The paper analyses the elements and their relations in data warehouse multi-dimension schema and gives the definitions of schema version based-on relation model.At the end,the paper gives the implementation method on schema version maintenance.

Key words: data warehouse, multidimensional schema, schema version, version maintenance

摘要： 数据仓库的“非易失性”包括“事实数据非易失”和“模式数据非易失”两个方面,“时间”维度保证了事实数据的非易失性,而模式数据的非易失性需要在数据仓库中建立模式版本来实现。论文分析了数据仓库多维模式中应包含的主要元素及其关系,提出基于关系模型设计并存储数据仓库的模式版本,给出了模式版本的结构定义及其更新维护的实现方法。

关键词: 数据仓库, 多维模式, 模式版本, 版本维护