I am seeing that some scalar UDFs have slowed down (X10!) since we converted to 2005 2 months ago. Its is not stats or indexing - all of that have been rebuilt/refreshed since conversion several times. I know that UDFs are slow since they are a cursor in disguise, but I wouldnt think that they would be slower on 2005 for any reason. I just thought I would throw this out and see if anyone had a similar experience. I will be running this on a dev box and comparing plans etc, but I just wanted to thow this out there<img src='/community/emoticons/emotion-1.gif' alt='' /><br /><br />Michael <br />MCDBA<br /><br />"I have a lovely bunch of coconuts" - Benny Hill
Are you using SCHEMABINDING option for those UDFs? http://blogs.msdn.com/sqlprogrammability/archive/2006/05/12/596424.aspx fyi. Satya SKJ Microsoft SQL Server MVP Writer, Contributing Editor & Moderator http://www.SQL-Server-Performance.Com This posting is provided AS IS with no rights for the sake of knowledge sharing. The greatest discovery of my generation is that a human being can alter his life by altering his attitudes of mind.
Nope. didn't know I could<img src='/community/emoticons/emotion-1.gif' alt='' /> I am unsure if I can however. I need to look to see if these are ever dropped and recreated rather than truncated.<br /><br />Michael <br />MCDBA<br /><br />"I have a lovely bunch of coconuts" - Benny Hill
oops.. seems like it shouldnt be used if the udf is tied to a table, which it is... Mike Michael MCDBA "I have a lovely bunch of coconuts" - Benny Hill
I;m bit unsure whether APPLY will help in this case, I believe more digging is needed in this case. Satya SKJ Microsoft SQL Server MVP Writer, Contributing Editor & Moderator http://www.SQL-Server-Performance.Com This posting is provided AS IS with no rights for the sake of knowledge sharing. The greatest discovery of my generation is that a human being can alter his life by altering his attitudes of mind.