By Belo, O.; Gomes, C.; Oliveira, B.; Marques, R.; Santos, V.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
ETL conceptual modeling is a very important activity in any data warehousing system project implementation. Owning a high-level system representation allowing for a clear identification of the main parts of a data warehousing system is clearly a great advantage, especially in early stages of design and development. However, the effort to model conceptually an ETL system rarely is properly rewarded. Translating ETL conceptual models directly into something that saves work and time on the concrete implementation of the system process it would be, in fact, a great help. In this paper we present and discuss a hybrid approach to this problem, combining the simplicity of interpretation and power of expression of BPMN on ETL systems conceptualization with the use of ETL patterns to produce automatically an ETL skeleton, a first prototype system, which has the ability to be executed in a commercial ETL tool like Kettle.