I'm all for normalizing data... but if no one is going to use the normalized form, then you shouldn't bother. In this case, it actually provides a negative ROI...

Heh... Like you say... touche`. You're absolutely correct about...

You bet... thanks for the feedback, Timothy.

Because, contrary to popular belief, Clustered Indexes are NOT totally contained in the data at the Leaf level... there is an Intermediate Level containing index rows AND a Root Level,...

So... how many columns?

cliffb (11/6/2008)

---

Jeff Moden (11/6/2008)

---

Cliff, take a look at the text file attached in the first post of this thread... not much to "pattern match" on... certainly, not enough to...

Thanks... and that's a very appropriate use for that code. If it was a huge table with large gaps, I was gonna suggest a different way.

David Jackson (11/7/2008)

---

Excellent. 😀

I particularly like the date formatting trick. Why didn't I think of that! :w00t:

Many thanks Hugo. I'll run the extract later today and post some...

I'm not...

Any bet's on that? 😉

USE TempDB

GO

--===== Create and populate a 1,000,000 row test table.

...

To make a really long story shorter...

Some manager was complaining about one of his developers as being non-compliant, difficult to control, and generally an arrogant moron. After many suggestions...

rbarryyoung (11/7/2008)

---

Local variables are entirely local in scope and visibility: whereas dynamic SQL makes a new batch scope for every invocation that will not be able to see the original...

Then, I have to agree with kshitij... unless I'm missing something, those all sound like "read" operations and there's no need for the overhead of building a 1+ million row...

Cliff, take a look at the text file attached in the first post of this thread... not much to "pattern match" on... certainly, not enough to build a whole...

No problem... could you provide some test data in a readily consumable format like I asked in my previous post? See the link in my signature for how to...