Efficient processing of spatial joins using R-trees

  • Authors:

  • Thomas Brinkhoff;Hans-Peter Kriegel;Bernhard Seeger

  • Affiliations:

  • Institute of Computer Science, University of Munich, Leopoldstr. 11 B, W-8000 München 40, Germany;Institute of Computer Science, University of Munich, Leopoldstr. 11 B, W-8000 München 40, Germany;Institute of Computer Science, University of Munich, Leopoldstr. 11 B, W-8000 München 40, Germany

  • Venue:
  • SIGMOD '93 Proceedings of the 1993 ACM SIGMOD international conference on Management of data
  • Year:
  • 1993

Quantified Score

Hi-index 0.00

Visualization

Abstract

Spatial joins are one of the most important operations for combining spatial objects of several relations. The efficient processing of a spatial join is extremely important since its execution time is superlinear in the number of spatial objects of the participating relations, and this number of objects may be very high. In this paper, we present a first detailed study of spatial join processing using R-trees, particularly R*-trees. R-trees are very suitable for supporting spatial queries and the R*-tree is one of the most efficient members of the R-tree family. Starting from a straightforward approach, we present several techniques for improving its execution time with respect to both, CPU- and I/O-time. Eventually, we end up with an algorithm whose total execution time is improved over the first approach by an order of magnitude. Using a buffer of reasonable size, I/O-time is almost optimal, i.e. it almost corresponds to the time for reading each required page of the relations exactly once. The performance of the various approaches is investigated in an experimental performance comparison where several large data sets from real applications are used.