Initial Optimization Techniques for the Cube Algebra Query Language: The Relational Model as a Target

Initial Optimization Techniques for the Cube Algebra Query Language: The Relational Model as a Target

Thomas Mercieca, Joseph G. Vella, Kevin Vella
Copyright: © 2022 |Volume: 18 |Issue: 1 |Pages: 17
ISSN: 1548-3924|EISSN: 1548-3932|EISBN13: 9781799893684|DOI: 10.4018/IJDWM.299016
Cite Article Cite Article

MLA

Mercieca, Thomas, et al. "Initial Optimization Techniques for the Cube Algebra Query Language: The Relational Model as a Target." IJDWM vol.18, no.1 2022: pp.1-17. http://doi.org/10.4018/IJDWM.299016

APA

Mercieca, T., Vella, J. G., & Vella, K. (2022). Initial Optimization Techniques for the Cube Algebra Query Language: The Relational Model as a Target. International Journal of Data Warehousing and Mining (IJDWM), 18(1), 1-17. http://doi.org/10.4018/IJDWM.299016

Chicago

Mercieca, Thomas, Joseph G. Vella, and Kevin Vella. "Initial Optimization Techniques for the Cube Algebra Query Language: The Relational Model as a Target," International Journal of Data Warehousing and Mining (IJDWM) 18, no.1: 1-17. http://doi.org/10.4018/IJDWM.299016

Export Reference

Mendeley
Favorite Full-Issue Download

Abstract

A common model used in addressing today's overwhelming amounts of data is the OLAP Cube. The OLAP community has proposed several cube algebras, although a standard has still not been nominated. This study focuses on a recent addition to the cube algebras: the user-centric Cube Algebra Query Language (CAQL). The study aims to explore the optimization potential of this algebra by applying logical rewriting inspired by classic relational algebra and parallelism. The lack of standard algebra is often cited as a problem in such discussions. Thus, the significance of this work is that of strengthening the position of this algebra within the OLAP algebras by addressing implementation details. The modern open-source PostgreSQL relational engine is used to encode the CAQL abstraction. A query workload based on a well-known dataset is adopted, and CAQL and SQL implementations are compared. Finally, the quality of the query created is evaluated through the observed performance characteristics of the query. Results show strong improvements over the baseline case of the unoptimized query.