A performance evaluation of storing XML data in relational database management systems

XML is an emerging standard for the representation and exchange of Internet data. Along with document type definition (DTD), XML permits the execution of a collection of queries, using XPath to identify data in XML documents. In this paper we examine how XML data can be stored and queried using a standard relational database management system (RDBMS). For this, we propose a technique for automatic mapping from an XML document to relations within the RDBMS. We demonstrate that our novel approach preserves the nested structure of the XML documents. By hiding database details we devise a seamless, transparent framework for user access to XML data. In order to achieve this, we propose a novel mechanism for translating an XPath query into an SQL statement. Furthermore, we propose efficient techniques for the construction of an XML document on the fly from the result set of the SQL statement. We also present findings in terms of query response time on the comparative performance of different techniques for the construction of an XML document on the fly.