XML to Relational Conversion Using Theory of Regular Tree Grammars

Abstract

In this paper, we study the different steps of translation from XML to relational models, while maintaining semantic constraints. Our work is based on the theory of regular tree grammars, which provides a useful formal framework for understanding various aspects of XML schema languages. We first study two normal form representations for regular tree grammars. The first normal form representation, called NF1, is used in the two scenarios: (a) Several document validation algorithms use the NF1 representation as the first step in the validation process for efficiency reasons, and (b) NF1 representation can be used to check whether a given schema satisfies the structural constraints imposed by the schema language. The second normal form representation, called NF2, forms the basis for conversion of a set of type definitions in a schema language L ¹ that supports union types (e.g., XML-Schema), to a schema language L ² that does not support union types (e.g., SQL), and is used as the first step in our XML to relational conversion algorithm.

This author is partially supported by NSF grants 00861 16, 0085773, 9817773.

We differentiate two terms - XML schema(s) and XML-Schema. The former is a general term for schema in XML model while the latter refers in particular to the XML schema language proposed by W3C [17].