Semantic Query Optimization in Object-Oriented Databases

In many relational database systems using a relational algebra-based query language, query optimization involves the syntactic modiflcation of queries into a \canonical form"(Dat90), and then choosing from possibly several methods of evaluating the query. Semantic query optimiza- tion (SQO) is the idea of semantically transforming a query using additional schema informa- tion, such as integrity constraints. The query is passed through three difierent phases of logical transformation: standardization, simpliflcation, and amelioration. The result of this process is a query that may, in fact, appear quite difierent from the original query posed by the user, but is guaranteed to return the same results given the same database instance. In this paper, we give a brief overview of the concepts of semantic query optimization, and present a survey of the literature. Most of the literature involves the application of SQO to relational systems, or to logic-based models such as Datalog(CGT89, Ull88, Ull89). To show how these techniques may be used with Object-Oriented Database Management Systems (OODBMS), we attempt to apply Chakravarthy et al.'s(CGM90) idea of semantically constrained axioms to an extension of Datalog called Complex-Prolog(GR89). Complex-Prolog supports the notions of object identity, classes, and inheritence. Finally, we present a summary and give some research directions for the application of semantic query optimization to OODBMS.