Modelando bases de datos multidimensionales

DWQ
Foundations of Data Warehouse Quality
National Technical University of Athens (NTUA) Informatik V & Lehr- und Forschungsgebiet Theoretische Informatik (RWTH) Institute National de Recherche en Informatique et en Automatique (INRIA) Deutsche Forschungszentrum für künstliche Intelligenz (DFKI) University of Rome «La Sapienza» (Uniroma) Istituto per la Ricerca Scientifica e Tecnologica (IRST)M. S. Hacid, U. Sattler Modeling Multidimensional Databases: A formal object-centered approach

Proc. of the Sixth European Conference on Information Systems 1998 (ECIS98).

DWQ : ESPRIT Long Term Research Project, No 22469
Contact Person : Prof. Yannis Vassiliou, National Technical University of Athens, 15773 Zographou, GREECE Tel +30-1-772-2526 FAX: +30-1-772-2527, e-mail: yv@cs.ntua.grMODELING MULTIDIMENSIONAL DATABASES: A FORMAL OBJECT-CENTERED APPROACH
Mohand-Said Hacid Ulrike Sattler

LISI-INSA LuFg Theoretical Computer Science 20, av. A. Einstein RWTH Aachen, Ahornstrasse 55 69621 Villeurbanne, France 52074 Aachen, Germany msh@lisi.insa-lyon.fr uli@cantor.informatik.rwth-aachen.de

ABSTRACT
In this paper, we propose a formal framework based on description logicsas a basis for both modeling multidimensional databases at the conceptual level and understanding related reasoning problems and services. Description logics are well-understood, object-centered formalisms which have already proved their suitability as a unifying framework for object-centered representation formalisms. We extend a description logic with new constructors which allow to defineoperators on cubes. The constructors we propose correspond in part to the operators on data cubes given in (Agrawal&al 1997).

1. INTRODUCTION
Recently, much attention has been focused on multidimensional databases. Many commercial systems based on their own data models have been developed. The main strength of multidimensional databases is their ability to view, analyze, and consolidate huge amountsof data. To do so, instead of presenting tables to the user, data is presented in the shape of so called datacubes which can be manipulated by using operators to cut out pieces from large cubes, change the granularity of dimensions, turn cubes, etc. Because of these functionalities, multidimensional databases play an important role in Decision Support Systems, for On Line Analytical Processing,and in Data Warehousing. In general, multidimensional databases provide two categories of tools: 1) Tools for integration, efficient storage and retrieval of large volumes of data. 2) Tools for viewing and analyzing data from different perspectives. These tools allow interactive querying of data and their analysis, often referred to by ''navigation'' since this way of querying seems much moreintuitive than classical ad-hoc queries.

For describing the interdependence of data in a multidimensional database, data cubes are regarded as an appropriate data model. In a data cube, each axis is associated with a dimension (e.g., time, space, or products) and its according values. Then, points in the cube (called cells) can be associated with values (called measures) of additional dimensions(like sales volumes). Displaying information to the user as well as navigation in this information are tasks that are heavily supported by this data model. In order to consolidate and analyse huge amounts of data, aggregation and roll-up play an important role. Data is often summarized at various levels of granularity and on various combinations of attributes. Therefore, queries are more complex andmay take long time to complete. To face up to this problem, a great deal of effort has been invested in developing techniques for optimizing queries involving views and aggregate functions (Harinarayan&al 1996, Gray&al 1996, Ho&al 1997, Levy&al 1996, Ross&al 1994). Unfortunately, a unifying framework for multidimensional databases is still missing. Hence, different multidimensional models,...