Exploiting vague queries to collect data from heterogeneous XML sources

This chapter describes a framework for querying heterogeneous XML data sources, that extends previous approaches for approximate query evaluation, by providing techniques for combining partial answers coming from different sources. This approach does not rely on a global schema shared by the sources, but it automatically adapts the query to the available data, providing the user with the XML elements satisfying the query to a certain extent. Based on this framework, a query language is described which allows the collection of as much information as possible from several heterogeneous XML sources. An algorithm for approximately evaluating a query on a single source and a strategy to join partial results coming from different sources are provided. Finally, an experimental validation of the approach in a peer-to-peer application scenario is presented. © 2010 Springer-Verlag Berlin Heidelberg.