The Research of Methodology in Models Mapping for ETL Processes Based on Model Driven

It is very important to optimize pattern design of ETL scenario´s to data warehouse´s integration in the ETL (extract-transformation-load) tools development, a good modeling process is helpful to the accuracy of description in data source, and makes the ETL workflow´s design more reasonable. To find out a methodology for facilitating the modeling and management of ETL processes, aimed at Alkis Simitsis and Panos Vassiliadis´ models, based on the formal definition of objects in meta-models, an improved design model and an optimization algorithm in mapping conceptual to logical models were proposed. After the improvement, the algorithm was able to mapping the simple source or the multi-source models in the foundation of ETL workflow. The experiment proved that this method realized the design of ETL based on model driven and provided the good foundation in the data integration by using an unified way to describe the data model.