The Rugby World Cup is much more exciting a prospect!

Last point: I have just looked at the what-about execution plan. Here (with 500 set as the limit, and no clustered index) the optimizer chooses a much better plan....

I think we can all agree on that one. In fact,...

That's just my view - I may be wrong about it. Never any need to apologise.

It's 7:30am and I need to...

RBarryYoung (3/1/2010)

---

True, but I prefer to try to get the optimizer to "do the right...

Ah, good, right. I see. Cool.

Yes worker time is CPU time spent by a SQL Server worker thread on a particular task, in the RUNNING state (as opposed...

RBarryYoung (3/1/2010)

---

Hmm, I thought that it was that it was doing the larger index first, when it should do the smaller first. But I guess that doesn't mean as...

Still a risk of the optimizer choosing a merge join with that one, Barry.

The text file in the ZIP is empty 🙁

The problem in your execution plan was the merge join. A poor choice by the optimizer.

Paul

Please...try this:

WITH Numbered

AS (

SELECT CF.ID,

...

agustingarzon (3/1/2010)

---

The LOB data is not binary data, just long strings, I'm attaching the table create script. There are 3 large fields (nvarchar-max) that can add up to 100kb in...

Maybe it's a collation issue. I normally blame collation issues :laugh:

A very balanced reply there Lutz. 😎

Huge thanks again for your awesome article review!

Alvin Ramard (3/1/2010)

---

I was labouring under a misapprehension as far as the Olympics were concerned:...

I think a concrete example might help illustrate the reasons for my discomfort with logical reads:

Consider the cost of a KEY lookup versus an RID lookup.

In the case of the...