Viewing 15 posts - 5,701 through 5,715 (of 8,415 total)
GilaMonster (3/3/2010)
http://www.sqlservercentral.com/Forums/Topic875858-338-1.aspx
Probably a variant of parameter sniffing, from all that's been said. The procs and exec plans have been made...
March 3, 2010 at 8:39 am
GilaMonster (3/3/2010)
March 3, 2010 at 8:26 am
david.pelizzari (3/2/2010)
March 3, 2010 at 8:21 am
RBarryYoung (3/2/2010)
And you know, this would all be a lot easier and simpler if you did this in a set-based way, instead of using a CURSOR and WHILE loops.
This is...
March 3, 2010 at 8:17 am
Jeff Moden
The exception to the rule, of course, would be if "TableWithIdentity" were a Temp table.
...and the exception to that would be if it were a global temp table 😀
Seriously,...
March 3, 2010 at 8:13 am
Jeff Moden
No doubt the CLR wins.
I only mention it because no-one else ever does 😛
Jeff Moden
March 3, 2010 at 8:10 am
Jeff Moden
Don't rely on the execution plan for differences in performance either.
A skilled and experienced person can tell more about performance from an 'actual' execution plan than profiler will ever...
March 3, 2010 at 8:04 am
Steve Jones - Editor (3/2/2010)
March 3, 2010 at 5:58 am
Barry-193141 (3/3/2010)
This way the newid() function only needs to run on an already randomized...
March 3, 2010 at 1:12 am
Jeff Moden (3/2/2010)
If you keep track of the pennies, you can probably afford to buy a new red stapler. 😉
That would be awesome! I can never find a stapler...
March 2, 2010 at 8:56 pm
Performance results for that (huge) thread were summarized very nicely by Florian Reischl on his blog:
http://florianreischl.blogspot.com/2009_09_01_archive.html
It covers all the main methods, including WHILE loops, Tally tables, XML, and CLR.
Paul
Spoiler: CLR...
March 2, 2010 at 8:53 pm
Seems like a good plan to me. It kinda bugs me to start blogging too...maybe.
March 2, 2010 at 8:44 pm
mstjean (3/2/2010)
Are you saying it's possible for the following statement to insert non-sequential values for the identity:
INSERT INTO TableWithIdentity SELECT TOP 100 * FROM SomeTable
Non-sequential/non-contiguous.
I feel ok commenting for Jeff...
March 2, 2010 at 8:18 pm
RBarryYoung (3/2/2010)
Personally, I am about convinced that this is an Optimizer bug of some kind, and that it is also the direct cause of the bad plan in this case.
It's...
March 2, 2010 at 8:06 pm
Viewing 15 posts - 5,701 through 5,715 (of 8,415 total)