一种基于间接索引桶的OLAP查询算法

A query algorithm for online analytical processing queries evaluation based on mediate index bucket

提高多表连接和聚集操作性能是OLAP查询中的关键问题之一。本文提出了一种基于间接索引桶的OLAP分组聚集查询算法MIBGA。该算法将维层次编码和事实表标识符分组集合进行有效结合，用间接索引桶代替目前流行的位图连接索引，并通过分组属性位图的位操作方式来快速完成OLAP查询。分析表明，该方法压缩了索引的存储空间，减少了I／O开销，有效地提高了多表连接的查询效率。

How to improve the performance of joining and aggregation is a key problem for OLAP query processing. This paper proposes a query algorithm MIBGA （Grouping and Aggregation based on Mediate Index Bucket）. The algorithm efficiently integrates fact table grouping sets into dimension hierarchical encoding, at the same time, substitutes mediate index bucket for popular bitmap joining index. The ultimate grouping and aggregation is completed by bit operation of grouping-attribute bitmap. As a result, this algorithm not only can greatly reduce index space and disk I/Os, but also can highly improve the efficiency of OLAP queries.