Thanks for the great feedback - especially the thing about slashes in a batch file!

Always frustrating when you don't solve the problem...but hey.

Ha. He'll just come back and say he was after the sum...

I don't go to the desert. I make phone calls to people that like to hurt other people :w00t:

stefan.gustafsson 60897 (3/18/2010)

---

Of course there is some difference between 72 bytes per row and 16 bytes per row.

What I am saying is that this difference is irrelevant for this problem....

So that answers the original question, albeit somewhat trivially.

The hard-coded value N'{00000000-0000-0000-0000-000000000000}' allows SQL Server to use the statistics to realize that the value specified falls completely outside the range...

Australia?

Thanks, that is great. It is 4am here now, so I will take a fresh look in the morning.

Thanks for the feedback!

It is NVARCHAR(50).

Do you really think there is no difference between indexing a GUID at 16 bytes per...

vidya_pande (3/18/2010)

---

Yes collate clause is required if user database collation is different than server collation.

Tempdb database follows the same collation as server collation. If some query operation uses tempdb...

SQL Server is using a different (and better) method of allocating memory. See http://blogs.msdn.com/psssql/archive/2009/09/11/fun-with-locked-pages-awe-task-manager-and-the-working-set.aspx for details.

Sorry, I didn't see this...

Now you are just being silly.

I'm going for a short break to prevent pork-chop Armageddon.

http://www.sqlservercentral.com/Forums/FindPost885514.aspx

stefan.gustafsson 60897 (3/18/2010)

---

The reason your example worked was not that you used UNIQUEIDENTIFIER instead of string, it was that you had a unique index and you had no duplicates in...