Evolution of Schema of XML-documents Stored in a Relational Database

XML is today a standard for manipulating semistructured data. One of widely used industrial solutions, especially for systems with a fairly well defined data structure, is storing XML in a relational database, while XML queries are converted to SQL queries to the underlying relational database. A software product that produces XML “interface” to an underlying relational database commonly requires revision of XML and relational schemas with every new version of the product. Those schema and subsequent data transformations are selected and performed using ad-hoc algorithms. We propose a framework, namely formal evolution model, allowing semi-automatic schema transformation and data transformation for a new product version, thus discarding the necessity of ad-hoc algorithm design. The framework also allows a-priory estimating of query conversion performance.