A Temporal Join Based on Rank Key-to-Address Transformation

Abstract

In this paper, we present an efficient nested-loop like algorithm for the temporal join operation in temporal relational databases. The algorithm is based on the concept of rank KAT (key-to-address-transformation) which was introduced by Ghosh in 1977. By pre-storing appropriate ranking information of endpoints of time intervals in the joining temporal relations, our algorithm can determine the resulting relation of a temporal join rapidly without requiring any scan on the joining relations. We further show that, for a given tuple in the outer relation satisfy the local predicates, the presented algorithm to determine all the inner relation tuples which match the given tuple is O(r+3/4), where r is the number of distinct endpoints of time intervals in the joining relations.