Techniques for Design and Implementation of Efficient Spatial Access Methods

In order to handle spatial data efficiently, aa required in computer aided design and g-data ap plications, a database management system (DBMS) needs an access method that will help it retrieve data items quickly according to their spatial location. In this paper we present a classification of existing spatial access methods and show that they use one of the following three techniques: clipping, overlapping regions, and transformation. From a practical point of view we provide a tool box supporting simple design of a spatial access method for a given point access method using one of the above techniques. We analyze the technique of transformation in more detail and show that our new concept of asymmetric partitioning is more retrieval efficient than the traditional symmetric approach. Furthermore we suggest a hybrid method combining the techniques of overlapping regions and transformation and provide an analysis and comparison of our new method. For data for which an analysis of Rand R+-trees was available, these comparisons demonstrate a superiority of our scheme.