A materialized view selection technique for XPath queries
Kazutaka Furuse, Keiichi Sato, Hanxiong Chen, Nobuo Ohbo · 2008
This paper proposes an effective method for selecting XPath views which contribute to improve query processing performance when they are materialized. This method finds a set of efficient materialized views from a given set of workload queries written in the XPath query language. The selected views are computed in advance and stored in databases so that they can be used to process XPath queries efficiently. The proposed method introduces a new data structure called view inclusiveness graph. This data structure enables us to apply the traditional greedy algorithm for finding effective views. Because of the powerfulness of the data structure, the proposed method can handle various XPath operators including wildcards, relative paths, and predicates. The proposed method has the advantage that it can take the cost of recomputing views and cost of storing views into consideration in the process of view selection. The experimental results with synthetic data sets show that the proposed method is effective and potential. Key–Words: Materialized views, View selection, XML, XPath, Greedy algorithm