Algebraic Operators for Processing Sets of Temporal Intervals in Relational Databases

Abstract

The efficient management of temporal data has become increasingly important for many database applications. Most commercial systems already allow the management of temporal data but the operational support for processing this data is still rather limited. One particular reason is that many extension proposals typically require considerable modifications of the underlying database engine. In this paper, we propose a lightweight solution where temporal operators are realized using a library of user-defined functions. This way the complexity of temporal queries can be drastically reduced leading to more readable and less error-prone code without touching the database system. Our experiments show that the proposed operators significantly outperform temporal queries formulated in pure SQL. In addition, we investigate the possibility to incorporate algebraic optimization strategies directly into our operator definitions which allow for further performance improvements.