Temporal issues in data warehouse systems

Abstract

With the wide acceptance of the data warehousing technology, data models and development methodologies are being proposed by researchers. While they recognize time as an important dimension, the suggested modeling is largely ad hoc and experience driven. On the other hand, research into temporal databases since the mid-eighties has made many fundamental contributions, and has created a formal and powerful framework both at modeling and technology levels. In this paper, we review the temporal issues in data warehousing, and propose a temporal model for data warehouses. The concepts of states, events, object life-cycles, temporal constraints and temporal operators are used to enrich the data model for warehousing. We propose a logical data warehouse design step that takes into account temporal characteristics of data, followed by a physical design step to address efficient storage and access. We also propose an integrated framework of applications and data warehouses, where the applications are made 'warehouse aware' for pushing complete and consistent data into the data warehouse. We also identify some important issues for future investigations.