OLAP Analysis Operators for Multi-State Data Warehouses

OLAP Analysis Operators for Multi-State Data Warehouses

Franck Ravat (Toulouse Institute of Computer Science Research - Toulouse I Capitole University, Toulouse, France), Jiefu Song (Toulouse Institute of Computer Science Research - Toulouse I Capitole University, Toulouse, France) and Olivier Teste (Toulouse Institute of Computer Science Research - University of Toulouse-Jean Jaurès, Toulouse, France)
Copyright: © 2016 |Pages: 34
DOI: 10.4018/IJDWM.2016100102
OnDemand PDF Download:
No Current Special Offers


Data reduction in Multidimensional Data Warehouses (MDWs) allows increasing the efficiency of analysis and facilitating decision-makers' tasks. In this paper, the authors model a MDW containing reduced data through a set of states. Each state is valid for a certain period of time; it contains only useful information according to decision-makers' needs. In order to carry out analyses in a MDW composed of multiple states, an extension of traditional OLAP analysis operators is required. In this paper, the authors define a set of OLAP operators compatible with reduced MDWs. For each operator, they propose a user-oriented definition along with an algorithmic translation. To show the feasibility and the efficiency of the proposed concepts, they implement the analysis operators in an R-OLAP framework.
Article Preview

1. Introduction

Multidimensional Data Warehouses (MDWs) organize data in a multidimensional way in order to support On-Line Analytical Processing (OLAP) analyses. A MDW schema is based on facts (analysis subjects) and dimensions (analysis axes). The facts contain analysis indicators, while a dimension includes analysis parameters organized in hierarchies from the lower granularity (most detailed) to the higher granularity (most general). In a classical MDW, all data are permanently stored and new data are periodically added. The increasing volume of MDWs makes the tasks of decision-makers more difficult since they may be lost during their analyses. On the other hand, information is usually timely sensitive; most of detailed information loses its value over time (Ravat & Teste, 2000). Nevertheless, data at high granularity levels are more stable, and it can generally fulfill decision-makers’ needs when analyses are carried out over older data (Skyt, Jensen, & Pedersen, 2008). For instance, an analyst may have interest in analyzing sale amounts by product’s brand for the last five years. However, as many of today’s brands did not exist before, the brand granularity level may be useless for an older period. As a result, the analyst may have no more interest in analyzing sale amounts by brand over the last ten years but by a higher and more stable granularity level, such as product’s category.

Facing large volumes of data among which a great amount of inadequate detailed data are found, our aim is to both increase the analysis efficiency and facilitate the analysts’ tasks. To this end, we propose a solution for supporting OLAP analyses over only relevant data over time. Firstly, we describe a conceptual MDW model based on data reduction to aggregate and then remove useless detailed data. Secondly, we propose a set of algebraic OLAP operators to support analyses over reduced data. The execution process of each operator is illustrated with the help of an algorithm. At last, we implement the reduced MDW model and the set of analysis operators in an R-OLAP framework. The framework aims at proving the feasibility of proposed concepts and evaluating the efficiency of carrying out analyses over reduced data.

This paper is composed as follows. Section 2 studies related works. Section 3 presents preliminary concepts of a reduced MDW illustrated with a case study. Section 4 describes our modeling solution for OLAP analysis operators. Section 5 presents a multi-state analysis framework showing the feasibility of our solution. Section 6 provides results of experimental assessments on the efficiency of the proposed OLAP operators.

Complete Article List

Search this Journal:
Open Access Articles
Volume 17: 4 Issues (2021): Forthcoming, Available for Pre-Order
Volume 16: 4 Issues (2020)
Volume 15: 4 Issues (2019)
Volume 14: 4 Issues (2018)
Volume 13: 4 Issues (2017)
Volume 12: 4 Issues (2016)
Volume 11: 4 Issues (2015)
Volume 10: 4 Issues (2014)
Volume 9: 4 Issues (2013)
Volume 8: 4 Issues (2012)
Volume 7: 4 Issues (2011)
Volume 6: 4 Issues (2010)
Volume 5: 4 Issues (2009)
Volume 4: 4 Issues (2008)
Volume 3: 4 Issues (2007)
Volume 2: 4 Issues (2006)
Volume 1: 4 Issues (2005)
View Complete Journal Contents Listing