Efficiently supporting order in XML query processing

Abstract

Query processing over XML data sources has emerged as a popular topic. XML is an ordered data model and XQuery expressions return results that have a well-defined order. However little work on how order is supported in XML query processing has been done to date. In this paper we study the challenges related to handling order in the XML context, namely challenges imposed by the XML data model, by the variety of distinct XML operators and by incremental view maintenance. We have proposed an efficient solution that addresses these issues. We use a key encoding for XML nodes that supports both node identity and node order. We have designed order encoding rules based on the XML algebraic query execution data model and on node encodings that does not require any actual sorting for intermediate results during execution. Our approach supports more efficient incremental view maintenance as it makes most XML operators distributive with respect to bag union. Our approach is implemented in the context of Rainbow [25], an XML data management system developed at WPI. We prove the correctness of our order encoding approach, namely that it ensures order handling for query processing and for view maintenance. We also show, through experiments, that the overhead of maintaining order in our approach is indeed neglectible.