Semantic Query Optimization on Object-Oriented Database

摘要

In this paper we discuss the problem of query optimization on object-oriented database. Rich semantic knowledge in terms of class hierarchy, aggregation hierarchy, and integrity constraints are considered in this work. In the effort of exploring the quantification of query performance, we develop a theorem to compute the maximum number of access in checking a predicate involving a nested attribute. From its corollary, we find that the decreased number of access of objects in earlier classes of the path have more effect on the total number of the access of objects under the assumption that all objects of all classes are uniformly distributed among all values of all attributes. Using this fact, we suggest that the predicate with a path of shorter length should be evaluated prior to another predicate with a path of longer length, if one path of the two is a sub path of the other. In the process of exploration, we conclude some guided rules and define efficiency ratio as measurement for quantifying query performance. Finally, a simple and easy-to-use algorithm is suggested for rewriting queries in order to achieve higher performance.