一种基于YFilter的XML数据流查询的改进算法

An extended algorithm based YFilter of XML stream querying

利用XML文档中各路径之间相似的特点,只扫描一次XML文档,把重复的XML标记所生成的事件组合成一个聚合事件,并在基于共享前缀的NFA算法YFilter基础上,提出一种改进算法PolYFilter,实现了聚合事件的谓词计算.实验表明,与YFilter相比,PolYFilter算法减少了大量相同的有限自动机状态转移的中间状态,避免了状态集的重复计算.特别是当XML文档比较大,且重复标记比较多的时候,PolYFilter有较好的过滤性能.

There are similarities between the various paths in the XML document,so an improved algorithm PolYFilter is proposed based on YFilter of NFA algorithm of shared prefix.PolYFilter combines all of the same events into a polymeric event running through the XML parser at most once,and it also implement evaluating predicate on polymeric event.The experiment shows that the improved PolYFilter can reduce numbers of operations in the set of intermediate states.Particularly,the algorithm is efficient to deal with large XML documents that have more similar events.