Modélisation multidimensionnelle de documents XML centrés-données
2010; Taylor & Francis; Volume: 19; Issue: 3 Linguagem: Inglês
10.3166/jds.19.313-345
ISSN2116-7052
AutoresYasser Hachaichi, Jamel Feki, Hanêne Ben‐Abdallah,
Tópico(s)Advanced Computational Techniques and Applications
ResumoToday’s international nature of commerce forced the opening of corporal information systems (IS) to accept data required for both transactional and decisional processes. To ensure the interoperability of so opened IS, XML has been adopted as a de facto standard format for exchanged data. In order to accommodate this data format in the design of decision support systems, we propose in this paper an automatic method for the design of data mart schemas from XML documents. The proposed method has the merit of automatically identifying all multidimensional elements, classifying them in terms of their analytical potential, and tracing them to the source which facilitates the definition of ETL procedures. For this, our method analyzes the DTD of the source documents and consults a sample of XML documents to determine typing information when needed. It relies on a set of identification rules that are independent of the domain of the XML documents. Furthermore, being automatic, it is supported by a CASE toolset that facilitates its exploitation.
Referência(s)