A Uniform Approach to Processing Temporal Queries

Research in temporal databases has mainly focused on defining temporal data models by extending existing models, and developing access structures for temporal data. Little has been done on temporal query processing and optimization. In this paper, we propose a uniform framework for processing temporal queries, which builds upon well-understood techniques for processing non-temporal queries. We start with an object-oriented model, and rely on its rich type system to model complex temporal information. The same query language is used to express temporal and non-temporal queries uniformly. A major benefit to this approach is that temporal query processing can be smoothly extended from an existing (non-temporal) query processing framework. For the purpose of query processing, we describe an object algebra, into which queries are compiled. Since the object algebra resembles the relational algebra, familiar relational query optimization techniques can be used. However, since the physical representation of temporal data and access methods differ from those of nontemporal data, new algorithms must be developed to evaluate the algebraic operators. We demonstrate that temporal queries can be processed and optimized under the existing query processing framework.