A Semantic Approach to ETL Technologies

Data warehouse architectures rely on extraction, transformation and loading (ETL) processes for the creation of anupdated, consistent and materialized view of a set of data sources. In this paper, we aim to support these processes byproposing a tool for the semi-automatic definition of inter-attribute semantic mappings and transformation functions.The tool is based on semantic analysis of the schemas for the mapping definitions amongst the data sources and thedata warehouse, and on a set of clustering techniques for defining transformation functions homogenizing data comingfrom multiple sources. Our proposal couples and extends the functionalities of two previously developed systems: theMOMIS integration system and the RELEVANT data analysis system.