Abstract
Spatial join is a fundamental operation for many spatial queries in Geographical Information Systems (GIS). Therefore, the query optimizer of a GIS needs to evaluate the selectivity of spatial joins, in order to find the best execution plan for a given query. This situation has made it necessary to find good and efficient estimators for spatial join selectivity. In particular, spatial join estimation with respect to sets of rectangles is necessary. Indeed, in GIS sets of rectangles are generated in order to produce a synthetic representation of real geometric values through the Minimum Bounding Rectangles (MBR).
Several methods for this estimation have been proposed in literature. One of the best methods is based on precalculated histograms, that describe the distribution of rectangles in the reference space using grid based data structures. The size of an histogram for a given dataset can be comparable to the size of the R-tree built on the same dataset [4].
In this paper we present a new technique for estimating spatial join selectivity considering sets of rectangles as datasets. In particular, we propose a technique that is independent of the distribution of the rectangles in the reference space and produces an auxiliary structure which is an order of magnitude smaller than the corresponding histogram. Indeed, the proposed technique is based on very few statistical parameters and on a unique grid shared by all datasets.
