Labeling XML nodes in RDBMS

Abstract

Relational Database storage for XML data is the most affordable and available solution. Meanwhile supporting document order and structure for dynamic XML data in RDBMS still needs more work. We present a new labeling approach for nodes in XML documents, which we call XMask, where each element is identified by two values (1) node Id (2) and node Mask. XMask can efficiently maintain the document structure and order with a short-length label that can fit in a 32-bit integer. We performed experimental evaluation between our approach and an interval approach model using XML data (XRel). Our approach outperformed the interval approach for complex queries. © Springer-Verlag Berlin Heidelberg 2006.