Overview of Dynamic Query Evaluationin Intensional Query Optimization

In 1], We introduced a new query optimization framework called inten-sional query optimization (IQO). This framework enables existing optimization techniques to be applied to queries over views that employ unions. Advanced database technologies and applications such as federation and mediation over heterogeneous database sources, object oriented databases and query languages, and data warehousing for decision support readily lead to such complex view deenitions, and thus tend to incur complex, expensive queries. The IQO framework provides that some of the query's expansions, also called unfoldings, may be explicitly \separated out" of the query. It deenes the notion of a discounted query, which is the query annotated to the eeect that some of its expansions have been removed. (These removed expansions are called the unfoldings-to-discount.) Thus this ooers us a way to recast complex queries, which use unions, into pieces for which we may have good optimizations. The query's answer set is then equivalent to the union of the answer sets of the unfoldings-to-discount plus the answer set of the discounted query. For this to be a viable overall optimization strategy, it must be possible to evaluate a discounted query less expensively than the query itself; indeed less expensively than the query itself less the cost of evaluating each unfolding-to-discount. We contend this is possible, and we motivate and sketch an eecient evaluation strategy for discounted queries. Consider the following example. Let there be two views in the DB: an employee relation, via the union of the ta relation and projections from the faculty and staa relations; and a beneets relation, via the union of projections from the life ins and health plan relations. Deene query Q to ask for the names of all employees of the physical plant, p p, whose beneets are provided by hmo: