有选择条件的并行连接改进算法

The improved parallel join algorithm for selective condition

针对有选择条件的连接易于造成选择率偏斜和重分布偏斜的问题 ,通过在各个处理结点中的A D(At tribute Distribute)临时关系以及主控结点中的A RD(Attribute Redistribute)临时关系 ,提出了并行连接改进算法 ,并且从理论上对该算法的空间复杂度进行了分析 .

To avoid selection skew and redistribution skew in joins where selections existed, the improved parallel join algorithms were discussed through the A-D (Attribute-distribute) temporary relations in the processing nodes and an A-RD (Attribute-Redistribute) temporary relation in a control node. The theoretical analysis of the spatial complexity of the algorithm was presented.