数据仓库系统中渐变维度更新策略的研究

Study on the strategies of updating dimensions in data warehouse system

数据仓库通常按多维模型组织数据,随时间的推移,维度不可避免地会发生变化,此时,需要及时对维度进行更新,以免影响数据仓库中数据的准确性。首先从理论上总结了数据仓库系统中维度更新的经典策略。针对复杂需求,需要混合使用基本渐变维度来实现更新,同时提出了一种将清除旧数据和保存历史相结合的方法,可进一步提高决策分析的高效性。

Data is always organized as multi-dimensional model in data warehouse,but dimensions are inevitably changed over time.In order to avoid impacting the data accuracy in data warehouse,it is necessity to update the dimensions in time.This paper summarizes the classical strategies of dimensions updating in data warehouse system at first.Aiming at complex requirements,hybrid SCDs need to be used to implement the dimensions updating.In order to improve the performance of OLAP,a new strategy combining the old data deleting and history track keeping is proposed.