XML数据流上的有序XPath查询处理.pdfVIP

计算机研究与发展，卷43(增刊)：464－470，2006,11. (第23届中国数据库学术会议，广州.) XML 数据流上的有序 XPath 查询处理 1 1 1 1 1,2 谢敏 王小锋 张新 孟小峰 周军锋 1 (中国人民大学信息学院，北京 100872) 2 (燕山大学计算机系，河北 秦皇岛 066004) (xiemin@) 摘要 XML数据流上的查询处理是最近研究工作的一个热点，如何高效地处理XML数据流上的XPath查询是其中的核心问题。之 前的相关工作主要考虑了无序XPath查询处理的情况，而在股票信息监控，新闻信息订阅等很多的XML数据流应用中常常需要 对有序XPath查询进行有效的支持。对于有序XPath查询的处理，之前的方法需要将查询进行分解，然后通过连接将分解后的 子查询得到的中间结果合并。针对有序XPath查询自身的特点，本文提出了在查询树上引入顺序和位置标记，记录查询结点之 间的顺序关系，并在此基础上提出了一种创新的XML数据流上的XPath查询处理算法OrderedXP。相比之前的工作，OrderedXP 能够大量地减少缓存的中间结果数目，而且不需要分解原来的查询，避免了额外的连接操作。详细的实验数据验证了OrderedXP 能够显著地提高有序XPath查询在XML数据流上的执行效率。 关键字 XML数据流；XPath；查询处理 中图法分类号 TP391 Ordered XPath Query Processing on XML Stream 1 1 1 1 1,2 Xie Min , Wang Xiaofeng , Zhang Xin , Meng Xiaofeng , and Zhou Junfeng 1(Information School, Renmin University, Beijing 100872) 2(College of Computer Science, Yanshan University, HeBei Qinghuangdao 066004) Abstract Recently, query processing on the XML stream is a hot topic in research community, in which how to efficiently handle XPath query on the XML stream is a core problem. On dealing with this problem, the previous work mainly concerns how to efficiently handle unordered XPath query, but in the application of XML stream like stock information monitoring, news feed monitoring and etc., we often need to handle ordered XPath query. On concerning these requests, the previous methods often break the XPath into some query Fragments and execute them separately, at the last stage these algorithms refer to an explicitly join to get the final results. On observing the characteristic of the ordered XPath query, we bring order specification and position specification into the query tree which may record the order relationship between the query nodes, and then we propos