This is one reason that I bear the pain of having IMPLICIT_TRANSACTIONs turned on in my SSMS settings.

It is nice to be able to issue a ROLLBACK after something like...

Ow. This thread hurts my eyes and my ears! 😛

So, having read the requirements a bit more thoroughly:Whistling: here's a quick hack which seems to work:

ALTER TRIGGER [dbo].[trig_test_insert]

ON [dbo].[cnsmr_accnt]

INSTEAD OF INSERT

AS

BEGIN

IF@@ROWCOUNT = 0 RETURN;

SET NOCOUNT ON;

SET ROWCOUNT 0;

;WITHUnioned AS

(

SELECTi.cnsmr_id,

i.cnsmr_accnt_id,

i.case_feed

FROMinserted...

Carl,

So I've read the thread a little more thoroughly now and hacked together something that appears to work.

See the original thread.

Paul

Cool. ...

An actual execution plan would be much more useful than estimated plans (we could see the run-time distribution of data across the threads for example).

In the posted plans, the PIVOT...

Carl Federl (6/24/2009)

---

Does anyone have an idea why row_number() in a trigger does not appear to work ? The exact same code in a batch works fine (substituting a...

For the record, the script Carl posted runs fine and produces the expected output on my 2005 and 2008 servers.

(that was the one saying ROW_NUMBER might not be working in...

Oh I see.

I guess I started using NULL because it somehow emphasises the...

Carl Federl (6/24/2009)

---

One of the issues that is discussed in the video was the database option to automatically update statistics, which does a sample, and stored procedure plans. If...

Hey Flo,

Nope - no difference at all, it just depends on my mood 🙂

ROW_NUMBER() OVER (ORDER BY (SELECT -42.7))

...works just as well.

Paul

That would do it!

🙂

Sure.

In some cases, one would want to add RECOMPILE to the procedure.

In others, just to one or more statements in the procedure.

In still others, to the procedure call itself.

I guess...

Perhaps because he was told to?

Just sayin'.

😀

Awesome 🙂