Extract Transfer Load Considerations of Temporal Data Sources to Data Warehouses

Abstract

Data warehouses are databases which store integrated enterprise-level information from information system data sources across the organization. These data sources may employ different database management systems, use different data formats, and use different value coding standards. Integrated data availability at the enterprise level from such diverse data sources requires transfers from transactional data sources, transforms them to the standard format, and loads them into the data warehouse. This process is known as the Extract Transform Load (ETL) process. At present, temporal databases which support validtime are available and temporal extensions to SQL are standard features since SQL 2011. The challenge is the temporal feature of temporal databases is different from the time dimension of conventional data warehouses. Thus specialized ETL procedures are required in the case that temporal databases are used as the data sources of data warehouses. In this paper, ETL process considerations for temporal databases sources are presented.