¡@

A view mechanism can provide a user with an appropriate portion of a database through data filtering and aggregation. Views are often materialized for query performance improvement, and in that case, their consistency needs to be maintained against updates of the underlying data. They can be either recomputed or incrementally refreshed by reflecting only the relevant updates. With the emergence of XML as the standard for data exchange on the Web, active research is under way on efficient storing and querying of XML documents with the DBMS. In this paper, we investigate the materialization of XML views and their incremental refresh for the case of a restricted class of views. The object-relational DBMS is employed to store XML documents and their materialized views, and the update log is used for deferred view refresh. Algorithms for checking a logged update¡¦s relevance to a view and for generating the optimized SQL statements to refresh the materialized view stored in an object-relational database through scanning of the update log are described. Experimental results show that our approach can be very effective in providing views of a large-scale XML warehouse on the Web.

Keywords: XML, materialized view, deferred incremental view refresh, Web, semistructured data

Retrieve PDF document (200511_01.pdf)

Received February 13, 2003; revised July 30 & October 13, 2004; accepted January 12, 2005.
Communicated by Ming-Syan Chen.
^* This Work was supported by the Basic Research Program of the Korea Science and Engineering Foundation, grant No. R01-2003-000-10395-0.