数据仓库跨版本透明查询系统的设计

Design of Cross-versions Transparent Query System in Data Warehouse

针对传统数据仓库系统中多维模式进化历史的挥发性问题,提出用版本元数据来记录数据仓库进化过程中的每一种多维模式状态,给出版本元数据结构,设计了跨版本透明查询系统及相应的查询算法。查询分解算法将用户基于一种模式结构提出的查询请求分解为在各个数据仓库版本上计算的子查询,集成算法将子查询结果进行必要的汇总和转换。

Aiming at the volatility of schema evolution histories in traditional data warehouse system, this paper proposes an idea that using version metadata keeps the multidimensional schemas history. In this paper, version metadata structure has been presented. Cross-versions transparent query system and corresponding algorithms are designed. The query split algorithm decomposes the query command proposed on one schema into several sub-queries on data warehouse versions, and the integration algorithm integrates the results of sub-queries by necessary gather and transform.