Multidimensional Tries Used for Associative Searching

A k-dimensional (kd) trie data structure is proposed for associative searching. It is similar to the kd tree, but is superior in the performance of searching and updating operations. The kd trie requires less time than the kd tree, and is smaller, using fewer secondary storage accesses. The kd trie is applicable to differential files used for the storage of updates to a static master file, and for filter files for multidimensional data. The top levels of the kd trie can serve as a filter file for the differential file, in determining subregions which have not been updated. Only in the worst case is the kd tree superior. Figures.