基于聚类网格文件的数据库操作并行联结算法  

Parallel Join Algorithms for Database Operations Based on IBGF files

作　　者：董隽[1] 王志海[1] 等 

机构地区：[1]北方交大人工智能研究所，北京市１０００４４

出　　处：《铁路计算机应用》1999年第6期10-12,共3页Railway Computer Application

摘　　要：对数据库运算有很多并行算法。由于联结运算开销大而且使用频繁,所以,联结运算的优化至关重要。作者提出了基于网格文件的两个新的并行算法,分别用于一致关系和下一致关系。因为这里给出的算法避免了当前常用的散列算法中必然的开销很大的划分阶段,所以能够提高联结运算的效率。There are a lot of investigating on parallel algorithms in database. Since join operation is frequently and commonly used, and it is a costly operation. , optimizing the method for performing join operation is very important. Two new parallel join algorithms are presented in this paper, in order to reduce the complexity of database operations.. These parallel algorithms, based on clustering and grid files, can improve the performance of the join operation by avoiding the expensive partitioning phase inherent in hash based parallel join algorithms. Simulations and results of the algorithms are also presented.

关 键 词：关系数据库 并行算法 多处理机 联结算法 聚类网格文件 

分 类 号：TP311.132[自动化与计算机技术—计算机软件与理论]