【作者】 杨利； 李霖； 昌月楼； 阳国贵；

【Author】 Yang Li Li Lin Chang Yue Lou Yang Guo Gui (Department of Computer Science,National University of Defense Technology,Changsha 410073)

【机构】 国防科技大学计算机系；

【摘要】 针对传统的位向量技术在改进连接算法执行效率时存在的不足,本文提出了一种对称位向量技术.在此基础上,以Hybrid-Hash连接算法为背景,提出了采用对称位向量技术和动态内外关系角色转换方法的启发式并行散列连接算法SPHHJ.利用本文开发的解析分析模型,给出了Hybrid-Hash连接算法、Hybrid-Hash连接算法加传统位向量过滤以及SPHHJ算法的性能比较.模拟结果表明,当两个参加连接的关系都很大且其规模接近时,SPHHJ算法的性能可以超过Hybrid.Hash算法加传统的位向量过滤.在其它情况下,SPHHJ算法退化为后者.更多还原

【Abstract】 To address the deficiency of traditional vector filtering,this paper proposes a new heuristric Hybrid-Hash join algorithm,SPHHJ,based on symmetric vector filtering technique and dynamic roles transformation between inner relation and outer relation.Under the guidance of the analytic model developed in the paper,the performance comparison between SPHHJ,Hybrid-Hash and Hybrid-Hash with traditional vector filtering,are given.The simulation shows that when the size of two join relations are close and large enough SPHHJ can outperform Hybrid-Hash with traditional bit vector filtering and that otherwise the former is same as the latter.更多还原