Maintaining Temporal Views Over Non-Historical Information Sources For Data Warehousing

Abstract

An important use of data warehousing is to provide temporal views ovthe history of source data that may itself be non-temporal. While recent work in view maintenance is applicable to data warehousing, only non-temporal views have been considered. In this paper, we introduce a framework for maintaining temporal views over non-temporal information sources in a data warehousing environment. We describe an architecture for the temporal data warehouse that automatically maintains temporal views over non-temporal source relations, and allows users to ask temporal queries using these views. Because of the dimension of time, a materialized temporal view may need to be updated not only when source relations change, but also as time advances. We present incremental techniques to maintain temporal views for both cases, and outline the implementation of our approach in the WHIPS warehousing prototype at Stanford