I didn't see veto, more like "was nervous about". Sounded like there was some wiggle room (my mgr would listen if I told him that it was a "fake...

without starting a thread war - you'll end up with an entirely different maintenance nightmare - the "update" nightmare. For example - if you wish to change the description...

That's a function of a clustered index. "In the case of a TIE, a unique number is associated to the key to keep the rows distinct" - that's the...

not sure I fully understand all of the criteria on qualifying sales, but here's a rough sketch.

The current best way to do this is to pre-generate the monthly and YTD...

p.buchsbaum (10/9/2007)

Really, when I change 1 line in your code, it's all over.

Create Clustered Index sales_ix On #Sales(Dt, Sales)

________In Place Of

Create Clustered Index sales_ix On #Sales(Dt, Salestext)

38 is...

There are a few options to look at. First thought is that you need to incorporate the month component into the grouping, or you will be generating one number...

This will do it in a scalar fashion. Something entirely different would need to be devised if you need to run it as a function.

http://blogs.msdn.com/bartd/archive/2007/07/26/a-simpler-way-to-converting-a-hexadecimal-string-to-varbinary.aspx

one sec - (quick back and forths going on - posting based on code a few answers up). there needs to be a case when then...

The compound cluster index is creating some odd sorting sequence. Apparently by having 2 data types, the key being stored internally is stored in some way I can't quite...

You may also need to throw in a "free" variable - to help it along. (declare @dumm with same type as @ord.). I ran into the same issue the...

you're missing a line:

Update T

Set @Ord = (Case When @Dt<>Dt Then 0 Else @Ord End) +1,

@Dt = Dt,

Ord = @Ord

From #TVen As T

WITH (INDEX(#TVen),tablock) -- The index here is...

Correct - missed that one. Since the disks are local in the case of mirroring, you wouldn't run into latency issues accessing SQL data on shared disks over remote...

This should probably give you what you want (didn't catch the specifics on all field names, so you will have to play find and replace). This is a shameless...

The server AND each database have their own collation settings. Could it be that that DB is set incorrectly?

Take a look at this over thread: http://www.sqlservercentral.com/Forums/Topic305998-266-1.aspx

If your data grows as fast you seem to be implying it will be (supermarkets tend to generate a LOT of transactions), then I'd tend to lean towards Enterprise version....