Assess Queries for Interactive Analysis of Data Cubes

Matteo Francia , Matteo Golfarelli , Patrick Marcel (1) , Stefano Rizzi , Panos Vassiliadis

1 BDTLN - Bases de données et traitement des langues naturelles

Assessment is the process of comparing the actual to the expected behavior of a business phenomenon and judging the outcome of the comparison. In this paper we propose assess, a novel querying operator that supports assessment based on the results of a query on a data cube. This operator requires (1) the specification of an OLAP query over a measure of a data cube, to define the target cube to be assessed; (2) the specification of a reference cube of comparison (benchmark), which represents the expected performance of the measure; (3) the specification of how to perform the comparison between the target cube and the benchmark, and (4) a labeling function that classifies the result of this comparison using a set of labels. After introducing an SQL-like syntax for our operator, we formally define its semantics in terms of a set of logical operators. To support the computation of assess we propose a basic plan as well as some optimization strategies, then we experimentally evaluate their performance using a prototype.

Domaines: Base de données [cs.DB]

Matteo Francia, Matteo Golfarelli, Patrick Marcel, Stefano Rizzi, Panos Vassiliadis. Assess Queries for Interactive Analysis of Data Cubes. 24th International Conference on Extending Database Technology (EDBT),, Mar 2021, Nicosia, Cyprus. ⟨10.5441/002/edbt.2021.12⟩. ⟨hal-03452663⟩