|
<snip> Use existing keys in your queries and you should be able to get sub-second response times. </snip> That is an approach I have been trying desparately to put in place here. Some of the files (read most...) have key fields that make absolutely no sense (considering that the value is ALWAYS the same...) the folks here don't use the field in order by clauses etc. Using the field since it is the first key increases performance by exponential factors, I can only assume that the developers here are afraid of typing an additional 6 characters..... Thanks, Tommy Holden -----Original Message----- From: midrange-l-bounces@xxxxxxxxxxxx [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of richard@xxxxxxxxxxx Sent: Wednesday, September 06, 2006 8:33 AM To: midrange-l@xxxxxxxxxxxx Subject: Re: Performance of ODBC vs. other access methods Here's a thought. In fact we've run into this with people building custom indexes over our WebDocs Document Management tables: Build logical files or SQL views where needed to support inquiries. Just make sure not to specify UNIQUE for any of the keys. As long as you don't specifiy UNIQUE on the keys, you can get the performance needed without causing problems. I don't think too many vendor packages rebuild tables on a regular basis any more. Hopefully :-) Then create a CL program that gets run each time you apply a vendor update that deletes your custom views and indexes and rebuilds them after the vendor update. Think of this as similar to the customization you would do after an OS/400 upgrade. I always advocate storing customizations in a CL program that gets run after the OS/400 upgrade. Short of that you just have to be very careful with testing when you write queries against your DB. Use existing keys in your queries and you should be able to get sub-second response times. Regards, Richard Schoen RJS Software Systems Inc. "Providing Your....iNFORMATION NOW!" Email: richard@xxxxxxxxxxxxxxx Web Site: http://www.rjssoftware.com Tel: (952) 898-3038 Fax: (952) 898-1781 Toll Free: (888) RJSSOFT ------------------------------ message: 9 date: Wed, 6 Sep 2006 08:27:05 -0500 from: "Holden Tommy" <Tommy.Holden@xxxxxxxxxxxxxxxxx> subject: RE: MIDRANGE-L Digest, Vol 5, Issue 1682 <snip> index the hell out of your database based on your average query types </snip> If only I could...in our shop our main software package is a 3rd party software & the mandate is to not change or customize the base product unless it's just a "have-to" case. Since adding indexes can sometimes have adverse effects on some base code it's just not always a viable solution (not arguing at all..In fact I support index usage whole-heartedly...) unfortunately some of us have to march to the beat of someone else's drum =( Thanks, Tommy Holden
As an Amazon Associate we earn from qualifying purchases.
This mailing list archive is Copyright 1997-2024 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].
Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.