Algorithms and applications for answering ranked queries using ranked views

Abstract

Ranked queries return the top objects of a database according to a preference function. We present and evaluate (experimentally and theoretically) a core algorithm that answers ranked queries in an efficient pipelined manner using materialized ranked views. We use and extend the core algorithm in the described PREFER and MERGE systems. PREFER precomputes a set of materialized views that provide guaranteed query performance. We present an algorithm that selects a near optimal set of views under space constraints. We also describe multiple optimizations and implementation aspects of the downloadable version of PREFER. Then we discuss MERGE, which operates at a metabroker and answers ranked queries by retrieving a minimal number of objects from sources that offer ranked queries. A speculative version of the pipelining algorithm is described.

DOI: 10.1007/s00778-003-0099-8

Extracted Key Phrases

25 Figures and Tables

Statistics

01020'04'05'06'07'08'09'10'11'12'13'14'15'16'17
Citations per Year

111 Citations

Semantic Scholar estimates that this publication has 111 citations based on the available data.

See our FAQ for additional information.

Cite this paper

@article{Hristidis2003AlgorithmsAA, title={Algorithms and applications for answering ranked queries using ranked views}, author={Vagelis Hristidis and Yannis Papakonstantinou}, journal={The VLDB Journal}, year={2003}, volume={13}, pages={49-70} }