TwigTable: Using Semantics in XML Twig Pattern Query Processing

In this paper, we demonstrate how the semantic information, such as value, property, object class and relationship between object classes in XML data impacts XML query processing. We show that the lack of using semantics causes different problems in value management and content search in existing approaches. Motivated on solving these problems, we propose a semantic approach for XML twig pattern query processing. In particular, we design TwigTable algorithm to incorporate property and value information into query processing. This information can be correctly discovered in any XML data. In addition, we propose three object-based optimization techniques to TwigTable. If more semantics of object classes are known in an XML document, we can process queries more efficiently with these semantic optimizations. Last, we show the benefits of our approach by a comprehensive experimental study.