Discover

Publisher Summary

DISCOVER operates on relational databases and facilitates information discovery on them by allowing its user to issue keyword queries without any knowledge of the database schema or of SQL. DISCOVER returns qualified joining networks of tuples, that is, sets of tuples that are associated because they join on their primary and foreign keys and collectively contain all the keywords of the query. DISCOVER proceeds in two steps. First, the Candidate Network Generator generates all candidate networks of relations, that is, join expressions that generate the joining networks of tuples. Second, the Plan Generator builds plans for the efficient evaluation of the set of candidate networks, exploiting the opportunities to reuse common sub expressions of the candidate networks. Keyword search is the most popular information discovery method because the user does not need to know either a query language or the underlying structure of the data. The search engines available today provide keyword search on top of sets of documents. When the user provides a set of keywords, the search engine returns all documents that are associated with these keywords. Typically, two keywords and a document are associated when the keywords are contained in the document and their degree of associativity is often their distance from each other.