Feasibility of aggregates in time constrained queries

• Published in: Information systems (Oxford) Vol. 21; no. 7; pp. 595 - 614
• Main Authors: Jukic, N., Vrbsky, S.V.
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Ltd 01.11.1996; Elsevier Science
• Subjects: Aggregates; Applied sciences; Approximate Query Processing; Approximation; Computer science; Computer science; control theory; systems; Computer systems performance. Reliability; Exact sciences and technology; Information systems. Data bases; Memory organisation. Data processing; Monotonically Improving Answers; Performance Analysis; Software; Time Constrained Queries; Time Constrained Queries; Approximation; Aggregates; Monotonically Improving Answers; Performance Analysis; Approximate Query Processing; Aggregation; Query formulation; Performance analysis; Database query; Semantics; Monotonicity
• ISSN: 0306-4379; 1873-6076
• DOI: 10.1016/S0306-4379(96)00030-0

Approximate query processing has been introduced as a way of producing monotonically improving answers to set-valued relational database queries with timing constraints. The approximate query processor utilizes the potentiality of processing parts of a relation at a time, rather than processing the whole relation at once. In this paper we describe how the capability of handling aggregates is added to the previously presented approximate query processor. We identify the semantic information necessary for approximate processing of aggregate queries, and we present a mechanism for maintaining that semantic information. The I/O overhead required by the approximate aggregates is presented and analyzed through a series of simulations. The simulation results provide information about the factors that influence the performance of the approximate query processor when processing queries that include aggregates.