Efficiently Querying XML Documents Stored in RDBMS in the Presence of Dewey-Based Labeling Scheme

Abstract

Storing an XML document as a single record of type BLOB (or sequence of bytes) in RDBMS has become a widely used solution that reduces the complexity of reassembling the original document. However, shredding and indexing the XML document, using special labeling methods to recover the document order, is still required to efficiently support data-centric queries. The Dewey based labeling method has been considered to be the most suitable labeling technique to support dynamic XML documents. In this paper, we present a new space-efficient and easy to process Dewey-based labeling scheme. The new label structure, which is composed of two components: (Parent, Child) in Dewey format, would significantly improve the performance of XML queries that are based on parent-child and sibling relationships. Furthermore, we introduce an efficient alternative approach to navigate upwards the XML tree, which can be used to validate ancestor relationships. We report on an extensive experimental label length evaluation and performance tests between our approach and some recent Dewey based approaches using well-known XML benchmarks.