Viewing 15 posts - 301 through 315 (of 18,926 total)
Something like that.
If you want 7 days ago midnight you can use something like this instead
SET @Start = DATEADD(D, -7, DATEDIFF(D, 0, GETDATE()))
December 20, 2011 at 5:44 pm
Been there, done that. The only way you really learn something is by whakking around it for hours :hehe:.
December 20, 2011 at 5:41 pm
GSquared (12/20/2011)
Dev (12/20/2011)
December 20, 2011 at 5:30 pm
Then tweak the dlookup. Not sure enough of the syntax to help you out here.
December 20, 2011 at 9:17 am
The estimates are WAY off.
I'd try with not exists and make this the clustered index fz_senddat.
That should improve this immensely.
December 20, 2011 at 8:43 am
You still have those values in the combobox right? Just get them there with the column index and you'll be fine.
December 20, 2011 at 8:41 am
Actual plan has WAY more info about what's going on in that query. 1 of the most important is actual rows vs estimates.
December 20, 2011 at 8:39 am
This is an estimated plan, almost useless to debug performance.
You see a ton of operators because you are selecting from a view, not a table.
There's a timeout error in the...
December 20, 2011 at 7:08 am
That's the estimated plan, we need the actual plan.
December 20, 2011 at 6:58 am
+1 on the clustered index benefits, especially on such a large table.
We'd need the actual execution plan to help any further.
December 20, 2011 at 6:03 am
Wrong object, correct event.
December 20, 2011 at 5:55 am
Dev (12/19/2011)
Steve Jones - SSC Editor (12/19/2011)
GilaMonster (12/19/2011)
December 19, 2011 at 11:40 pm
December 19, 2011 at 10:44 pm
The other issues with views of views of views of views, is that you often access the same objects multiple times.
Now the optimizer is EXTREMELY smart. If it can...
December 19, 2011 at 9:22 pm
You need to rescript the whole thing and control it yourself. Here's how you can script a trace from the trace definition's tables.
December 19, 2011 at 8:54 pm
Viewing 15 posts - 301 through 315 (of 18,926 total)