|
|
|
@inproceedings{DBLP:conf/sigmod/HouOD91,
author = {Wen-Chi Hou and
Gultekin {\"O}zsoyoglu and
Erdogan Dogdu},
editor = {James Clifford and
Roger King},
title = {Error-Constraint COUNT Query Evaluation in Relational Databases},
booktitle = {Proceedings of the 1991 ACM SIGMOD International Conference on
Management of Data, Denver, Colorado, May 29-31, 1991},
publisher = {ACM Press},
year = {1991},
pages = {278-287},
ee = {http://doi.acm.org/10.1145/115790.115837, db/conf/sigmod/HouOD91.html},
crossref = {DBLP:conf/sigmod/91},
bibsource = {DBLP, http://dblp.uni-trier.de}
}
BibTeX
An error-constrained COUNT query in relational algebra is of the form "Estimate COUNT(E), where E is a relational algebra expression, such that the error in the estimate is at most e". The general approach is to evaluate an estimator for COUNT(E) using a large enough sample from input relations such that the error is less than the specified bound with a certain confidence level.
We present an approach which utilizes double sampling for error-constrained COUNT queries. We compare this approach with another approach, called adaptive sampling, in terms of the sample sizes needed to obtain the desired error bound with a given confidence level.
We have implemented both approaches (i.e., adaptive sampling and double sampling) in a prototype, real-time DBMS called CASE-MDB. We present some of the results of our performance evaluation experiments.
Copyright © 1991 by the ACM, Inc., used by permission. Permission to make digital or hard copies is granted provided that copies are not made or distributed for profit or direct commercial advantage, and that copies show this notice on the first page or initial screen of a display along with the full citation.
CDROM Version: Load the CDROM "Volume 1 Issue 2, SIGMOD '75-'92" and ...