A method of using semijoins to optimize queries with ADT functions

ADTs (A bstYYzct Data Types) have been known as a promising feature for making the relational database meet the requirements of CAD/CAM applications. This extension has brought a new dimension to query optimization. Based on the belief that the execution costs of selections and projects are inexpensive, conventional query optimization methods focus only on minimizing the execution costs of joins. However when a selection involving ADT functions(briefly, AD 2’ selection) is included in a query, the computation cost of the ADT functions may become an important factor to the totaJ execution cost of the query and therefore must be taken into consideration. The reason is that the evaluation of an ADT function is often computationally expensive. Therefore the conventional query optimization strategy is no longer suitable for such queries. We need a new query optimization strategy focusing on ADT selections.