一种找出较优重写查询的启发式算法  

A heuristic algorithm for finding out relatively optimal rewritten query

作　　者：陈长清[1] 程恳[2] 

机构地区：[1]华中科技大学软件学院,湖北武汉430074 [2]华中科技大学继续教育学院,湖北武汉430074

出　　处：《华中科技大学学报（自然科学版）》2005年第1期13-15,共3页Journal of Huazhong University of Science and Technology(Natural Science Edition)

摘　　要：提出了一种实用的启发式算法 .该算法使用了一种简单的代价模型来比较查询之间的相对代价 ,并利用多项式时间从实化视图得到的多个重写查询中找出一个执行代价较节省的重写查询 ,不需要查询优化器的支持 ,减少了优化器的负担 .实验结果表明该算法找出的较优重写查询的执行时间比穷尽搜索算法的略有增加 ,而搜索时间和空间显著减少 .利用该算法可快速地响应查询 ,能显著改进数据仓库的查询性能 .A heuristic algorithm was presented to find out an approximately optimal query from those rewritten ones by materialized views. A simple cost model was used to compare the relative expense of different queries, and a relatively optimal one with a polynomial time was located. It did not need the support of query optimizer and mitigated its burden. Experiments have shown that the execution time of the approximately optimal guery was just a little addition than that of the optimal query. The proposed algorithm can be used to answer queries fast and significantly improve data warehouse.

关 键 词：数据仓库 实化视图 重写查询 启发式算法 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]