Research On Predicate-based Query Processing Of XML Streams

With the XML data on the internet increasing rapidly, it is widely applied in the emerging application systems. In these applications, plenty of XPath expressions are processed against XML documents that arrive rapidly in stream form. Consequently, the major research challenge of query processing on XML stream is to efficiently handle a large number of XPath expressions.In this thesis, it is focused on that related to query processing on XML stream and efficiently handle a large number of XPath expressions on XML streams with focusing on the processing of simple XPath expressions and branched XPath expressions is researched. A novel predicate-based method is proposed for the processing of simple XPath expressions. The method supports the efficient processing of wildcard character and descendant axis of simple path expressions. The performance of the method are evaluated according to experiments and show that our predicate-based method takes less time than existing automaton-based method. To improve the filtering performance , a new data structure is introduced, predicate index implemented through multiple stages of hashtables. Next, all branched XPath expressions are decomposed into simple XPath expressions and the branched queries are evaluated based on the output of simple XPath expressions filtering. On the basis of these, the design of a model for XML stream query processing is presented and provide a detailed description of major function of its components.