Viewing 15 posts - 55,066 through 55,080 (of 59,072 total)
On a busy system, if it's a batch proc, wait 5 minutes after it's done, and it won't be cached any more.
On a busy system, if it's a GUI proc,...
--Jeff Moden
Change is inevitable... Change for the better is not.
September 11, 2007 at 11:00 pm
Somebody,
If I have a lot of candidates with good resumes all lined up, I'll ask some questions outside of what the job specifically requires as "tie-breakers".
For example... let's say I...
--Jeff Moden
Change is inevitable... Change for the better is not.
September 11, 2007 at 10:44 pm
Have you tried sp_HelpLogins? ![]()
--Jeff Moden
Change is inevitable... Change for the better is not.
September 11, 2007 at 9:27 pm
Subtracting 2 milliseconds from the date won't do it because dates have a 3.3 millisecond resolution.
And, instead of fooling around with the time for end date, why don't you...
--Jeff Moden
Change is inevitable... Change for the better is not.
September 11, 2007 at 9:09 pm
There is no time such as 24:00:00.000 because that would represent the next day ![]()
--Jeff Moden
Change is inevitable... Change for the better is not.
September 11, 2007 at 9:02 pm
Heh... a Yak by any other name... ![]()
--Jeff Moden
Change is inevitable... Change for the better is not.
September 11, 2007 at 8:53 pm
Sure... look at the problems you're having now with trying to keep the data unique. Which do you think will be more effecient... doing a two line insert (INSERT/SELECT UNION...
--Jeff Moden
Change is inevitable... Change for the better is not.
September 11, 2007 at 8:46 pm
Tk,
At the risk of ticking off "Mr Rogers", just some additional information... I believe that "performance tuning" effectiveness ratios are as follows:
1% - Hardware
2% - Server configuration
2% - Proper Indexing
95%...
--Jeff Moden
Change is inevitable... Change for the better is not.
September 11, 2007 at 8:38 pm
Finer resolution on locks will sometimes decrease performance... nothing faster than a full table lock because it is only one lock. Of course, you might not want to do that...
--Jeff Moden
Change is inevitable... Change for the better is not.
September 11, 2007 at 8:12 pm
That helps, but better to have the partitions on different physical disks for real performance gains...
--Jeff Moden
Change is inevitable... Change for the better is not.
September 11, 2007 at 8:08 pm
No need for a trigger... You can revoke permissions from individual users by column in a table or view. Lookup REVOKE in Books Online for the full syntax.... but here's a...
--Jeff Moden
Change is inevitable... Change for the better is not.
September 11, 2007 at 7:34 pm
It wasn't negative!!! It was friendly advise that you're probably commiting some form of SQL suicide by needing to do this!!! ![]()
There's only one...
--Jeff Moden
Change is inevitable... Change for the better is not.
September 11, 2007 at 7:19 pm
Is it a bug or did the boys in Redmond just decide to change a default like they did with the CONCATENATE NULL YIELDS NULL setting?
--Jeff Moden
Change is inevitable... Change for the better is not.
September 11, 2007 at 7:14 pm
The "system" creates them... As you found out, it only creates them when the first instance is needed whether created by the backup code or use of an IDE. Best...
--Jeff Moden
Change is inevitable... Change for the better is not.
September 11, 2007 at 7:06 pm
Won't have poor performance if it's for multi-row batches... on busy systems, most batch procs will probably need to recompile due to data changes in as little as 5 minutes. GUI code is another...
--Jeff Moden
Change is inevitable... Change for the better is not.
September 11, 2007 at 7:02 pm
Viewing 15 posts - 55,066 through 55,080 (of 59,072 total)