Towards semantic temporal support in data integration

Central idea of a data warehouse system is to provide an integrated, unified and analysis-oriented view on data of several operational and external information sources, e.g. transactional databases, applications or web data sources. Due to business demands, increasing amounts of data in these sources and data warehouses are stored with respect to time. Therefore, existing time information has to be utilized during data integration to gain consistent and correct integration results. Unfortunately, dealing with temporal data in a non-temporal manner, e.g. using non-temporal SQL, is complex and error-prone. This paper presents an approach for the integration of data in data warehouse systems which provides built-in and flexible support for handling temporal data in combination with different types of time information. The overall architectural framework is presented and concepts of a semantic temporal algebra for data integration are discussed in detail.