Incorporating an implicit time dimension into the relational model and algebra

Abstract

The prevailing approach to modeling the time dimension in databases is, in one form or another, the use of an explicit representation of time. In this paper we propose a temporal relational model and algebra, based on temporal semantics, to incorporate an implicit time dimension in databases. In temporal semantics, time-varying relations are an indexed collection of ordinary relations, one for each moment in time. A temporal database is modeled by a collection of time-varying relations. Temporal databases are queried using a temporal relational algebra (TRA) which extends the relational algebra point-wise upon the set of natural numbers. Although TRA lacks the ability to explicit manipulate time, we show how temporal aggregation and when-type queries can be formulated using a technique called tagging. The formal properties of TRA, which can be used for query optimization, are also outlined. We also compare our work with other proposed temporal algebras.