Updating XQuery Views Published over Relational Data
Ling Wang, Elke Angelika Rundensteiner · 2003
Abstract. Managing XML data using relational database systems, in-cluding query processing over virtual XML views that wrap relational sources, has been heavily studied in the last few years. Updating such virtual XML views, however, is not well studied, although it is essen-tial for building a viable full-featured XML data management systems. XML view update is a challenging problem because of having to ad-dress the mismatch between the two rather dierent data models and distinct query paradigms. In this paper, we tackle the XQuery view up-date problem, in particular, we focus on the round-trip XML view update subproblem. This case, characterized by a pair of loading and extraction mappings that load XML data into the relational store and extract ap-propriate XML views, is very common in practice, as many applications utilize a relational engine for XML document storage. We discuss and prove the updatability of such views. We also present a decomposition-based update translation strategy for solving this problem. As evidence of feasibility, we have implemented the proposed strategies within the Rainbow XQuery system. Experimental studies are also given to assess the performance characteristics of our update system in dierent scenar-ios. 1