Bjvrn You should aim for a key structure that avoids "overloading" a single key value with records and at the same time gives you a good hit rate.
Typical problem areas (though not yours in this case) are keys of 0 and 1, or with high numbers of null (empty) string entries. If you can post some example keys and selection criteria used we will see if we can optimise it for you. At the moment I suspect it will be a composite key manipulated in a virtual - but let's see some samples first. Regards JayJay -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of [EMAIL PROTECTED] Sent: 15 February 2006 17:27 To: [email protected] Subject: SV: [U2] [u2][ud] Poor select performance Hi JayJay, I guess you meant that I should create an index on the @id field. The keys look like this 66-0000112. I'm not sure if you say I should try: SELECT A_FILE WITH @ID <= '66-000001' AND @ID >= '66-9999999' PErhaps this is not what you mean since it's difficult to tell if 66-0000012 is bigger than 66-0000013. I've tried it but with less performance than my original i-descriptor index. Bjvrn Eklund ------- u2-users mailing list [email protected] To unsubscribe please visit http://listserver.u2ug.org/
