Consistently updating XML documents using incremental checks with Xqueries

by Kane, Bintou.

When updating a valid XML Data or Schema, an efficient yet light-weight mechanism is needed to determine if the update would invalidate the document. Towards this goal, we have developed a framework called SAXE. First, we analyzed the constraints expressed in XML schema specifications to establish constraint rules that must be observed when a schema or an XML data conforming to a given XML Schema is altered. We then classify the rules based on their relevancy for a given update case. That is, we show the minimal set of rules that must be checked to guarantee the safety for each update primitive. Next, we illustrate that this set of incremental constraint checks can be specified using generic XQuery expressions composed of three type of components. Safe updates for the XML data have the following components: (1) XML schema meta-queries to retrieve any constraint knowledge potentially relevant to the given update from the schema or XML data being altered, (2) retrieval of specific characteristics from the to-be-modified XML, and (3) an analysis of information collected about the XML schema and the affected XML document to determine validity of the update. For the safe schema alteration, the components are: (1) XML schema meta-queries to retrieve relevant information from the schema (2) analysis and usage of retrieved information to update the schema, and (3) propagation of the changes to the XML data when necessary. As a proof of concept, we have established a library of these generic XQuery constraint checks for the type-related XML constraints. The key idea of SAXE is to rewrite each XQuery update into a safe XML Query by extending it with appropriate constraint check subqueries. This enhanced XML update query can then safely be executed using any existing XQuery engine that supports updates - thus turning any update engine automatically into an incremental constraintcheck engine. In order to verify the feasibility of our approach, we have implemented a prototype system SAXE that generates safe XQuery updates. Our experimental evaluation assesses the overhead of rewriting as well as the performance of our loosely-coupled incremental constraint check approach compares with the more traditional first-changedocument and then revalidate-it approach. 2