Tree Pattern XML Queries Answering And Rewriting Using Materialized Views

With the continuous development of Internet, XML is playing a more and more important role in databases. The demand for XML queries grows constantly. Thus W3C has proposed XQuery language, as the standard XML query language. The relationship between XQuery and XML is similar to that between SQL and relational databases. XQuery combined the semi-structured feature of XML, can extract kinds of complicated query pattern from XML documents, and then refactor to new XML documents for users.Due to the semi-structure of XML data and the variety of XML queries, the efficiency of most XML queries can’t meet the requirements of applications. Using materialized views is an effective method to improve query efficiency. Many researches study the containment between materialized views and queries to reduce access to the source databases, so as to optimize the query. However, most researches only deal with the answering problem between simple XPath queries and materialized views and seldom support the flexible XML tree pattern query with multiple return nodes like XQuery.This paper proposes a tree pattern XML queries answering and rewriting algorithm, which is used to determine if an XML query can be answered by an XML tree pattern view, and then rewrite a compensation query for the views. The algorithm can handle the answering and rewriting problems between XPath queries, as well as XML tree pattern queries with multiple return nodes, which extends the application scope of this kind of answering algorithms. Experiments show the effectiveness of the algorithm.