Re: Index Files

[Arnold Foelster <afoelster@UMUC.EDU>]

Richard, thanks for the tip.

What is it considered too many rows to be stored in an array?

"Knapp, Richard" wrote:
>
> Maybe.  Keep in mind that indexes can be physically larger than the tables 
>that they index and for very large tables that can be a problem.
>
> For the kind of work that I do, network trips are the chief culprit in 
>slowing down the process.  SQR's ability to evaluate and process every little 
>atom of every little cursor can lead to being careless about how this 
>processing is done.  I find the notion of chunking to be useful here.  For 
>example, instead of setting up a cursor to bring back 2 or 3 columns 100,000 
>times and processing these as they arrive one by one, bring back the whole lot 
>and store them in an array which you can process later.  This makes much more 
>efficient use of the network packet size which means many fewer trips to the 
>server which means faster processing.
>
> Richard Knapp
> Database Programmer/Analyst
> Administrative Systems Project
> University of Missouri System
> 573-884-3157
> knappr@umsystem.edu
>
> -----Original Message-----
> From: Wayne C. Findley [mailto:wfindley@JPSCO.COM]
> Sent: Friday, June 07, 2002 5:41 PM
> To: SQR-USERS@list.iex.net
> Subject: Index Files
>
> Hi Everyone,
>
> If I create indexes to match every select statement in my SQR would the SQR
> run faster seeing that I'm working with mammoth tables?
>
> How would the SQR know to use the index?
>
> Anxious to get process moving faster.
>
> Help!
>
> Wayne