摘要： 结构连接的效率直接影响XML查询的性能。经典的Anc-Des-B+算法在判断双亲/孩子关系时跳过双亲节点的后裔(非孩子)节点的能力不强。为此，基于区间编码的思想提出一种改进的编码方法，把每个节点译码为六元组，并增加双亲节点的信息。给出的ZParent算法可以跳过孩子列表中所有不参与连接的元素节点，只需要扫描一次列表P和列表C，即可实现基于该编码的结构连接计算。实验结果表明，该方法具有较好的时间性能。

关键词: XML查询, 扩展区间编码, B+树索引, 结构连接

Abstract: The efficiency of structural join directly influences the function of the XML query. The classical Anc-Des-B+algorithm’s ability of skipping the descendant codes(not child) is not effective when judging the parent/child relationships. Based on the idea of extended region coding, this paper puts forward a kind of improved method which codes every node into a group of 6 elements including the parent’s information. It also puts forward the ZParent algorithm which can skip all nodes that don’t participate in join and can carry out join calculation with only one time to scan the P list and C list. Experimental result shows that the method has better time function.

Key words: XML query, extended region coding, B+ tree index, structural join

中图分类号: 

• TP311