Semantic transformations for XML queries

"The ever-increasing adoption of XML has created a need to ensure that XML query languages perform efficiently. Query optimization and transformation for XML query languages, both syntactically and semantically, have received much attention from research communities in recent years. However, due to the fast progress of the application of XML data management solutions, XML-Enabled Database Management Systems still face several challenges. Among these challenges is query processing, especially the processing of XML queries specified with XPath axes and redundancies that may exist in predicates used in XML queries. Semantic query optimization utilizes constraints in XML schemas to directly optimize a given query with a set of optimization rules. Due to the current complexity of the XML data structure which is enabled by rich semantics in XML Schemas, semantic query optimization should be performed in a more systematic manner. For a complete solution, this research proposes a series of semantic transformations to transform given XML queries to semantically equivalent, but more efficient, XML queries for optimization purposes, by using the semantics provided in XML Schemas. The proposed semantic transformations are grouped into three categories: (1) Semantic Path Transformations, (2) Semantic Transformations for XPath Queries Specified with Predicates, and (3) Semantic Transformations for XPath Queries Specified with XPath Axes. After a semantic transformation is applied to an XML query, the equivalent semantic XML query can be processed more efficiently by an XML data management system and returns the same result set. The proposed semantic transformations are then translated into a series of algorithms which are implemented and empirically evaluated for their efficiency and effectiveness. The experimental studies were carried out by using both real data sets (DBLP) and Benchmark data sets (Michigan) to illustrate that the majority of semantic transformations achieved significantly improved performance in XML query processing; this also enabled the research presented here to identify semantic transformations as optimization devices". -- Abstract.