maruf24 (8/3/2008)

---

i m hving a clusterd index on the column which cantain some duplicate records may be more than 1000 ....table contain more thn 10 millions of records....

when i...

Jeffrey Williams (8/3/2008)

---

I probably wouldn't use a dynamic SQL approach, because that is going to be much harder to work with on the report side anyways. How are you...

Jeffrey Williams (8/3/2008)

---

I would (personally) forget about any process that updates or tries to maintain this column using a trigger. I would just create it as a computed column...

Sorry, was in a hurry. I should have posted them (more than 1) without even being asked.

There are a couple of ways to do it... the method with the...

Probably not... it would probably recompile even if it were hard coded just because of data changes...

That will work just fine, but only if a single row is inserted at a time. If there's a chance that more than 1 row can be inserted at...

For what? The original code you posted?

If so, why not just insert the header information in a set based manner and then use that information to control another set...

Good point, Carl. Thanks.

Barry... you might want to look for the word "Cursor" in the above script... 😀 Heh, late nite posts get me, too! 😉

Other than that, nice conversion! 🙂

rbarryyoung...

Heh... Ummm... no... what's your user name again? 😉

http://members.iinet.net.au/~bofh/bofh/bofh2.html

http://members.iinet.net.au/~bofh/

The "SO" stands for "SysOP"... you can figure...

vkres1,

Forget about the identity column... if you want to store the data in the correct physical order, then put a clustered index on the column(s) that you want the data...

I don't think it will, especially if parallelism kicks in. If...

Try your code, Barry. 🙂

Ummm... you're missing the point a bit. The order in which something is inserted doesn't really matter in an RDBMS. Why do you think you need a "natural...