Optimizing Fixed-Schema XML to SQL Query Translation
Rajasekar Krishnamurthy, Raghav Kaushik, J.F. Naughton · 2002
Recently, there has been a lot of work on evaluating XML queries over data stored in rela-tional database systems. The vast majority of this work has focused on the cases where either the relational schema is not fixed (so the problem is to find a good relational schema for a given XML workload) or the XML schema is not fixed (so the problem is to develop generic strategies for exporting XML views of relational data). While these cases are interesting, in practice a third scenario, in which both the source relational and target XML schemas are fixed, seems highly relevant. We show that even in this highly constrained environment, there is a lot of freedom in the SQL that can be generated to evaluate a given XML query. Furthermore, we show through experiments with a commercial RDBMS that by exploiting the underlying relational constraints and the properties of a given XML to relational schema mapping, it is possible to generate SQL queries that perform an order of magnitude better than those generated by more naive translations. Motivated by these performance-enhancing opportunities, we present a constraint-aware XML to SQL query translation algorithm for path expression queries.