Advantages of complex SQL types in storing XML documents

This contribution looks at and evaluates new ways to map XML documents to relational database systems preserving their hierarchical structure. In particular we take advantage of the advanced features available following the SQL:2003 standard, which defines complex structure and collection types. Starting out with a survey of existing methods, basic questions concerning appropriate and efficient forms of mapping XML documents to SQL-compliant data types are addressed. This leads to a reversible conversion method, which is subsequently implemented in a prototype application. The design of the mapping is geared towards existing, mature relational database management systems (DBMS). It turns out that among the products considered, IBM Informix seems to offer the best support for complex structure and collection types in the context of our application. To better assess the performance of the mapping process, the observed running time and the required in-memory and database storage of the implementation are measured and evaluated.