A Survey on Temporal Data Warehousing

TLDR

Data warehouses are specialized repositories that support decision making, and because decision processes rely on historical trend analysis, effective time management becomes essential. This survey examines the range of temporal data warehousing challenges involved in accurately capturing how information evolves over time within data warehousing systems. The authors classify these challenges into four topics—data/schema changes in the warehouse, data/schema changes in the mart, temporal querying, and temporal warehouse design—and survey each, concluding with a discussion of open research issues and their implementation in commercial tools.

Abstract

Data warehouses are information repositories specialized in supporting decision making. Since the decisional process typically requires an analysis of historical trends, time and its management acquire a huge importance. In this paper we consider the variety of issues, often grouped under term temporal data warehousing, implied by the need for accurately describing how information changes over time in data warehousing systems. We recognize that, with reference to a three-levels architecture, these issues can be classified into some topics, namely: handling data/schema changes in the data warehouse, handling data/schema changes in the data mart, querying temporal data, and designing temporal data warehouses. After introducing the main concepts and terminology of temporal databases, we separately survey these topics. Finally, we discuss the open research issues also in connection with their implementation on commercial tools.