Improving the maintainability of data warehouse designs: modeling relationships between sources and user concepts

  • Authors:
  • Alejandro Maté;Juan Trujillo;Elisa de Gregorio;Il-Yeol Song

  • Affiliations:
  • University of Alicante, Alicante, Spain;University of Alicante, Alicante, Spain;University of Alicante, Alicante, Spain;Drexel University, Philadelphia, USA

  • Venue:
  • Proceedings of the fifteenth international workshop on Data warehousing and OLAP
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

In data warehouse (DW) development, a series of mappings must be specified between user concepts and data source elements, in order to identify which sources must undergo an integration process. Until now, these mappings are either assumed to be implied by name matching or identified according to the designer's experience. Then, the result is implemented as Extraction/Transformation/Loading (ETL) processes. Since ETL processes relate elements at the logical level, designers cannot adequately analyze how a change in requirements or in the data sources affects the analysis capabilities. Furthermore, this approach makes it difficult to perform incremental changes in DW design, requiring in some cases to perform the whole analysis again. In this paper we present a set of semantic mappings that relate user concepts specified by requirements to those obtained from data sources. In turn, this allows us to accurately identify how any potential change affects the different structures and ETL processes. As a DW evolves over time, our approach easily allows us to incorporate new concepts, as well as any change introduced at requirements or data sources into the DW repository with no need to redesign the whole DW. In order to show the application of our proposal, we show a real case study focusing on the Digital library of the University of Alicante.