Multi-Dimensional Joins

Abstract

We present three novel algorithms for performing multi-dimensional joins and an in-depth survey and analysis of a low-dimensional spatial join. The first algorithm, the Iterative Spatial Join, performs a spatial join on low-dimensional data and is based on a plane-sweep technique. As we show analytically and experimentally, the Iterative Spatial Join performs well when internal memory is limited, compared to competing methods. This suggests that the Iterative Spatial Join would be useful for very large data sets or in situations where internal memory is a shared resource and is therefore limited, such as with today's database engines which share internal memory amongst several queries. Furthermore, the performance of the Iterative Spatial Join is predictable and has no parameters which need to be tuned, unlike other algorithms.

The second algorithm, the Quickjoin algorithm, performs a higher-dimensional similarity join in which pairs of objects that lie within a certain distance epsilon of each other are reported. The Quickjoin algorithm overcomes drawbacks of competing methods, such as requiring embedding methods on the data first or using multi-dimensional indices, which limit the ability to discriminate between objects in each dimension, thereby degrading performance. A formal analysis is provided of the Quickjoin method, and experiments show that the Quickjoin method significantly outperforms competing methods.

The third algorithm adapts incremental join techniques to improve the speed of calculating the Hausdorff distance, which is used in applications such as image matching, image analysis, and surface approximations. The nearest neighbor incremental join technique for indices that are based on hierarchical containment use a priority queue of index node pairs and bounds on the distance values between pairs, both of which need to modified in order to calculate the Hausdorff distance. Results of experiments are described that confirm the performance improvement.

Finally, a survey is provided which instead of just summarizing the literature and presenting each technique in its entirety, describes distinct components of the different techniques, and each technique is decomposed into an overall framework for performing a spatial join.