Query processing algorithms for temporal intersection joins

Abstract

Intersection join processing in temporal relational databases is investigated. An analysis is presented of the characteristics and processing requirements of three types of intersection join operators: the time-join, temporal equijoin on the surrogate, and temporal equijoin on the temporal attribute. Based on the physical organization of the database and on the semantics of the operators, several algorithms were developed to process these joins efficiently. The primary cost variables were identified for each algorithm, and their performance is compared to that of a conventional nested-loop join procedure. It is shown that the algorithms developed can reduce processing costs significantly.