Dealing with Discrepancies in Wrapper Functionality

Overview & Synopsis

Much of the world's information is stored electronically in data sources. The data sources can be full-fledged databases, simple files, HTML pages or specialized data sources that possess diverse query processing capabilities. The common architecture to integrate such sources consists of {\it mediators} that give a global view over the content of all sources, and {\it wrappers} that give a local view of each source. Answering queries in this architecture is a difficult problem due to the wide range of capabilities of data sources. This paper presents a solution to this problem in the context of the \Duery processor. We provide a tool to the wrapper implementor to describe the capabilities of the wrapper in fine detail. When a wrapper is registered with the mediator, the mediator uploads the capabilities of the wrapper, and smoothly integrates these capabilities into query processing. Our solution is novel both in the level of detail permitted by the tool and its easy incorporation into existing query optimization strategies. In this paper we describe: the query processing of \Dthe language for specifying wrapper capabilities, the algorithms that integrates these capabilities into query processing, and an implementation of these techniques in the \Drototype.