Towards a Query Language for Multihierarchical XML: Revisiting XPath

Ionut E. Iacob, University of Kentucky
Alex Dekhtyar, University of Kentucky

Abstract

In recent years it has been argued that when XML encodings become complex, DOM trees are no longer adequate for query processing. Alternative representations of XML documents, such as multi-colored trees [7] have been proposed as a replacement for DOM trees for complex markup. In this paper we consider the use of Generalized Ordered-Descendant Directed Acyclic Graphs (GODDAGs) for the purpose of storing and querying complex document- centric XML. GODDAGs are designed to store multihierarchical XML markup over the shared PCDATA content. They support representation of overlapping markup, which otherwise cannot be represented easily in DOM. We describe how the semantics of XPath axes can be modified to define path expressions over GODDAG, and enhance it with the facilities to traverse and query overlapping markup. We provide efficient algorithms for axis evaluation over GODDAG and describe the implementation of the query processor based on our definitions and algorithms.

 

URL: http://digitalcommons.calpoly.edu/csse_fac/216