Heh... looks like you've been bitten before, eh Juan? 

So far as the ANSI_NULLs thing goes, the default setting, when you create an SQL Server instance, is ON.  Unless someone...

Cool... Thanks for the info, Robert.  I never tested to see if a trigger fired on a zero row insert... they way I write most of my triggers, it probably...

Guess you were the straw that broke the proverbial Camel's back this morning... have been fighting with some really stupid requirements written by some really stupid System's Analysts for one...

Horse feathers... asking "WHY" is very often the key to understanding the problem well enough to help you do what you need.

Since I can't ask "WHY" you want to do...

Yep, I agree with Serqiy about the denormalized thing... this is the wrong kind of update... and, if doing the concatenation in SELECTs turns out to be a chore, just add...

Dunno about Juan, but some folks can be pretty tough on even unrelated code.  Lookup some of Celko's replies for an example of what I mean... the intentions are good...

Doh!  I missed that part of the question.  Yes, Ron is correct but do remember what I said... It's session specific from SQL Server.

Thanks for posting your own solution... Lordy, I hope someone comes up with a different way... I really have bad gut feeling about updating system tables...

Not only do what SQLBill suggested, but why 23 indexes?  And, are any of them clustered?  Are any of them affected by the update?

That should work but I haven't tried it that way...

And, yep, the code I posted up to the "AS" was from Books OnLine... the rest is from some code...

PATH is session specific... and each use of xp_CmdShell creates a new session.  The only way to do this is to make sure that whatever code you are running contains...

Yep... that would be correct and that would also be why most folks think it doesn't work in the trigger as capturing the number of rows inserted, updated, or deleted...

Perfect... using the same function I used previously, adding quarter to this is easy...

DECLARE @StartDate DATETIME

DECLARE @EndDate   DATETIME

DECLARE @Days      INT

DECLARE @BinSize   INT

    SET @StartDate = '01/01/2006'...

I do it the same way that Lowell and Michael do to maintain the ability to use indexes if one exists.  The only difference is, I don't trust users even...

Juan,

Post your trigger code... we're only guessing until we see the code...