基于聚类的非一致性数据库查询重写  

Query Rewritings Based on Clusters in Inconsistent Databases

作　　者：谢东[1] 杨路明[1] 蒲保兴[1] 刘波[1] 

机构地区：[1]中南大学信息科学与工程学院,湖南长沙410083

出　　处：《小型微型计算机系统》2007年第12期2199-2202,共4页Journal of Chinese Computer Systems

基　　金：湖南省教育厅科研基金(05C671)资助;中南大学创新基金(ZB018)资助

摘　　要：在非一致性数据库上,以元组匹配技术所产生的聚类和概率数据库的元组概率为基础,提出了可信聚类概率和可重写查询判断方法.考虑了最普通的IC情况(key-to-key和nonkey-to-key),给出了无连接和有连接的查询重写方法.连接查询重写方法缩小了用于连接的中间结果集中可信聚类的元组数量,有效地提高了查询性能.实验使用TPC-H决策支持基准的数据和查询进行性能研究,分析了聚类基数和数据库尺寸等相关因素的影响,结果显示方法是有效的.This paper presents the probability of believable cluster and the decision method for rewritable queries in inconsistent databases, they are based on clusters that are produced by the tuple matching technology and tuple probabilities of probabilistic databases. The non-join and join query rewriting methods are proposed by considering the prevailing IC situations （key-to-key and nonkey-to-key）. The join query rewriting method condenses the number of tuples in believable clusters of middle result sets for join, and enhances the query performance effectively. The experiment presents a performance study using the data and queries of the TPC-H decision support benchmark and analyses some effects of relative factors such as the cluster cardinality and database size. The experiment results show that the method is effective.

关 键 词：关系数据库 非一致性数据库 查询重写 聚类概率 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]