Efficient Processing of Complex Twig Pattern Matching

As a de facto standard for information representation and exchange over the Internet, XML has been used extensively in many applications. And XML query technology has attracted more and more attention in data management research community. Standard XML query languages, e.g. XPath and XQuery, use twig pattern as a basic unit to match relevant fragments from a given XML document. However, in most existing work, only simple containment relationships are involved in the twig pattern, which makes it infeasible in many cases. In this paper, we extend the original twig pattern to complex twig pattern (CTP), which may contain ordered relationship between query nodes. We give a detailed analysis of the hard nuts that prevent us from finding an efficient solution for CTP matching, and then propose a novel holistic join algorithm, LBHJ, to handle the CTP efficiently and effectively. We show in experimental results that LBHJ can largely reduce the size of intermediate results and thus improve the query performance significantly according to various metrics when processing CTP with ordered axes.