Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ncar!boulder!ccncsu!handel.CS.ColoState.Edu!sweeneyd
From: sweeneyd@handel.CS.ColoState.Edu (daniel j sweeney)
Newsgroups: comp.databases
Subject: Re: optimising sql
Keywords: Cost Estimations for queries
Message-ID: <4113@ccncsu.ColoState.EDU>
Date: 12 Feb 90 03:48:25 GMT
References: <1990Jan31.010510.15242@welch.jhu.edu> <3601@dev.dtic.dla.mil>
Sender: news@ccncsu.ColoState.EDU
Reply-To: sweeneyd@handel.UUCP (daniel j sweeney)
Organization: Colorado State University, Fort Collins, CO  80523
Lines: 46

In article <3601@dev.dtic.dla.mil> jkrueger@dev.dtic.dla.mil (Jonathan Krueger) writes:
>francois@welch.jhu.edu (Francois Schiettecatte) writes:
>
>>Does anyone out there know of any books which might help
>>me to optimise sql statements which are to be sent to a 
>>database. What I really need are a set of guidelines, of
>>does and donts, of what type of query to use, etc.
>
>It isn't this simple.

    [some of the response deleted]

>Next get the right performance analysis tools.  At a minimum, you need
>a way to measure how expensive your queries are and a way to find out
>how your engine is executing them.  Use these tools.  You'll likely
>find that the queries you thought were problems can indeed be made to
>go faster, but that it won't help your application nearly as much as
>optimizing others you never considered.  This may lead you back to
>database design again.

    [some of the response deleted]

>Jonathan Krueger    jkrueger.dtic.dla.mil   uunet!dgis!jkrueger


I'm a student at CSU taking my first database class.  I recently read
and article in the ACM Transactions on Databases March 1989, entitled:

"On Estimating the Cardinality of the Projection of a Database Relation"
by Ahad, Rao, and McLeod.  The authors discuss their formula for 
estimating the cost of a query (in disk accesses) with exceptable error
results.  The formula does not require a scan of the relation.

This article has intrigued me.  I would like to do a project for my class
on estimating the cost of queries.  However, I am not familar with
what performance analysis tools are available or any schools doing
research in this area.  If anyone can point me towards articles, research
and software (for Vax, HP, Sequent Balance or any UNIX based OSs), I
would be grateful.

Thanks,

-- 
Dan Sweeney                            sweeneyd@handel.cs.colostate.edu
Department of Computer Science         sweeneyd%colostate.csnet@relay.cs.net 
Colorado State University (RAMS)       Fort Collins, CO 80523