Optimization of Hypothetical Queries in an OLAP Environment

Abstract

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

Explanation of exceptional values in multi-dimensional business databases

“…To the best of our knowledge, Balmin et al (2000) and Lakshmanan et al (2007) are the only published research works that address sensitivity analysis in OLAP databases in a significant way. In Balmin et al (2000), the authors have developed the SESAME system for the processing of hypothetical queries.…”

“…In Balmin et al (2000), the authors have developed the SESAME system for the processing of hypothetical queries. In many OLAP software products, sensitivity analysis is not possible at the moment.…”

Efficient intensional redefinition of aggregation hierarchies in multidimensional databases

“…A credulous approach leads to accept both alternatives, yielding alternative hierarchies, an undesired result. A skeptical approach, on the opposite, prevents grading loans from borrower b 3 . We have chosen the second approach.…”

“…However, the approach is strictly operational, and no priorities are imposed among alternatives. Studies of irregular hierarchies can be found in works on hypothetical queries for OLAP [3], and on multiple scenarios [13]. While these works deal with "what if" sort of queries, and do not consider rollups as functions, our work incorporate exceptions from the beginning with the concept of rollup function embedded.…”

UML-Based Modeling for What-If Analysis

“…[5] uses constraint formulae to create hypothetical scenarios for what-if analysis, while [9] explores the relationships between what-if analysis and multidimensional modeling. [4] presents the Sesame system for formulating and efficiently evaluating whatif queries on data warehouses; here, scenarios are defined as ordered sets of hypothetical modifications on multidimensional data. In all cases, no emphasis is placed on modeling and design issues.…”