Schema-free XQuery

  • Authors:

  • Yunyao Li;Cong Yu;H. V. Jagadish

  • Affiliations:

  • Department of EECS, University of Michigan, Ann Arbor, MI;Department of EECS, University of Michigan, Ann Arbor, MI;Department of EECS, University of Michigan, Ann Arbor, MI

  • Venue:
  • VLDB '04 Proceedings of the Thirtieth international conference on Very large data bases - Volume 30
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

The widespread adoption of XML holds out the promise that document structure can be exploited to specify precise database queries. However, the user may have only a limited knowledge of the XML structure, and hence may be unable to produce a correct XQuery, especially in the context of a heterogeneous information collection. The default is to use keyword-based search and we are all too familiar with how difficult it is to obtain precise answers by these means. We seek to address these problems by introducing the notion of Meaningful Lowest Common Ancestor Structure (MLCAS) for finding related nodes within an XML document. By automatically computing MLCAS and expanding ambiguous tag names, we add new functionality to XQuery and enable users to take full advantage of XQuery in querying XML data precisely and efficiently without requiring (perfect) knowledge of the document structure. Such a Schema-Free XQuery is potentially of value not just to casual users with partial knowledge of schema, but also to experts working in a data integration or data evolution context. In such a context, a schema-free query, once written, can be applied universally to multiple data sources that supply similar content under different schemas, and applied "forever" as these schemas evolve. Our experimental evaluation found that it was possible to express a wide variety of queries in a schema-free manner and have them return correct results over a broad diversity of schemas. Furthermore, the evaluation of a schema-free query is not expensive using a novel stack-based algorithm we develop for computing MLCAS: from 1 to 4 times the execution time of an equivalent schema-aware query.