【摘要】 如何减少联机分析处理中多表连接和压缩维属性连接关键字,对查询数据进行有效地分组聚集操作,成为联机分析处理查询处理的关键问题。为此,提出了一种基于多维层次编码的新型预聚集算法MDHEPA。该算法充分利用编码长度较小的多维层次编码及其前缀,对事实表中的数据进行快速地分组聚集计算,大大减少和简化了多表连接操作,提高了联机分析处理查询效率。理论分析和实验结果表明,该算法是有效的。更多还原

【Abstract】 How to reduce multi-table join,compress the dimension attribute join keywords and effectively aggregate the query data was the key issues for Online Analytical Processing(OLAP) query evaluation.To solve this problem,a novel pre-aggregation algorithm,MDHEPA(pre-aggregation based on the multi-dimension hierarchical encoding),was proposed.By using the small multi-dimension hierarchical encoding and their prefix paths,MDHEPA could rapidly aggregate the clustered fact data,so as to drastically reduce the multi-table join effort and completely remove one or more join operations.As a result,the algorithm could remarkably improve the efficiency of OLAP queries.The analytical and experimental results showed that the MDHEPA algorithm was more efficient than other existing algorithms.更多还原