Stars on steroids: Fast evaluation of multi-source star twig queries in path materialization-based XML databases

Despite a large body of work on XML twig query processing in relational environment, systematic study of XML join evaluation has received little attention in the literature. In this paper, we propose a novel and non-traditional technique for fast evaluation of multi-source star twig queries in a path materialization-based RDBMS. A multi-source star twig joins different XML documents on values in their nodes and the XQuery graph takes a star-shaped structure. Such queries are prevalent in several domains such as life sciences. Rather than following the conventional approach of generating one huge complex SQL query from a twig query, we translate a star query into a list of SQL sub-queries that only materializes minimal information of underlying XML subtrees as intermediate results. We have implemented this scheme on top of a path materialization-based XML storage system called SUCXENT++. Experiments carried out confirm that our proposed approach built on top of an off-the-shelf commercial RDBMS has excellent real-world performance.