Thanks for the data. Not sure if I'll need it, but I'm sure someone will.

I had a problem similar to this once when I forgot to tell the agent to use the DB Mail profile to send alerts. Is it possible you haven't...

I'm not sure what you mean by an online notification. Do you mean having data on a web page, or do you mean an e-mail, or something else?

The transaction log can't be eliminated. The transaction is (essentially) processed in there, then in the actual tables.

If that transaction expands the file, then there's no real way around...

The problem is probably the common misuse of "ie", with "eg" being what was intended. After all, "FROM" is a predicate, just like "WHERE", which would make the statement...

GilaMonster (8/27/2008)

---

Marios Philippopoulos (8/27/2008)

---

I'm considering a clustered index because of the GROUP BY and MAX clauses. Clustered indexes usually perform better in queries where ranges of rows are involved. Nonclustered...

I missed the TN column in the Where clause. That would be needed too.

Because of the B-tree nature of indexes, a non-clustered index should do just fine for this.

Here's...

Rather than relying on a shotgun approach like the missing index DMV, I tend to hit the low performance queries and tune them one at a time. That can...

Are you in Management Studio or Business Intelligence Development Studio?

If that query is the only thing being considered, how about both?

create clustered index CIX_Table on dbo.Table (CID, IDt)

But, as always, things like inserts, updates, deletes, and other...

You could build your own test system for that kind of thing, and it certainly could use AdventureWorks, but you're almost certainly better off getting something rather than building it.

Exactly.

You're welcome.

I guess it depends on the tests you're trying to do. What performance testing are you looking at doing?

Why build a function for it? That's row-by-row processing. The join I gave you should do what you need. I've used that type of construct many times...