The Research Of XML Query Processing Methods Based On Materialized Views

With the development of new Internet technology, Web services and data exchange applications prevail on Internet. Due to the large amount of heterogeneous data on the Web, a standard format of data representation and data exchange is badly needed. Fortunately, eXtensible Markup Language (XML) emerges timely, which is a standard data representing language proposed by W3C and is platform independent and self-describable. As XML is increasingly being used in data intensive applications, the number and the size of XML databases expand rapidly and thus efficiently managing XML data and processing XML queries have become the focus of database researchers. As in the context of relational databases, caching view results can significantly improve query performances. Therefore, it is vitally important to research on materialized XML view mechanism. Currently, only few works has addressed it. Crucial problems still required in-depth research, such as view selection for caching, view based XML query processing and incremental maintenance of XML views.In this paper, we propose methods of XML view selection, view_based XML query processing and incremental XML view maintenance. Document trees are used as XML data model and path expressions as view definition language. We mainly present view selection algorithm based on mining frequent query patterns, cache loading algorithm, view-based XML query processing algorithm and incremental XML maintenance algorithm for various update operations.Firstly, we propose a framework of selecting XML views for caching, including algorithms of enumerating frequent query sub-patterns, computing frequency of query sub-patterns and selection of materialized views for caching.Secondly, we propose a method for XML query processing based on cached view results. On the basis of the concept of query answerability, we present XPath and XQuery query processing algorithms using single view and multiple views. In addition, a cost model is built for evaluating the cost of query processing using materialized views and that of issuing queries to source database.Thirdly, we propose an incremental XML view maintenance strategy aiming...