Normalized Relational Storage for Extensible Markup Language (XML) Schema

Abstract

Problem statement: The use of XML as the common formats for representing, exchanging, storing, integrating and accessing data posses many new challenges to database systems. Most of application data are stored in relational databases due to its popularity and rich development experiences over it. Therefore, how to provide a proper mapping approach from XML model to relational model become the major research problems. Current techniques for managing XML in relational technology consider only the structure of an XML document and ignore its semantics as expressed by keys and functional dependencies. Approach: In this study we present an algorithm for generating an optimal design for XML in relational setting. The algorithm is based on computing a set of minimum covers for all functional dependencies on a universal relation when given XML Functional Dependencies (XFDs) and the schema information. However we need to deal with the hierarchical nature of XML and to define XFDs in this structure. Results: We show that our algorithm is efficient in terms of reducing data redundancy and preserving semantic expression. Conclusion/Recommendations: Being able to infer XML functional dependencies constraints to relational views of XML data is a first step towards establishing a connection between XML and its relational representation at the semantic level.