RBarryYoung (2/28/2010)

---

Worse, though the claim that the we should use CPU instead Logical IO is clearly stated both here, and in Joe's papers, no attempt at justification is made in...

Barry,

I seem to have struck a nerve, though that was not my intention. I simply found your posts a little verbose and hard to understand. The parts that...

You're welcome.

I have one question: why don't you use a linked server for this requirement?

You could then use the four-part name syntax, or EXECUTE...AT for pass-through queries.

Paul

Did you mention that size isn't everything? 😉

One would hope that such a monster was at the very least partitioned, and quite...

Well yes, quite.

As I...

lmu92 (2/28/2010)

---

Now that's what I call "significant improvement"!!!!

Seems like I need to look deeper into it, meaning to test it against a larger data set. Just to make sure the...

On the subject of logical reads as a performance metric.

I find Joe's reasoning to be sound - he presents his case lucidly, with plenty of examples and some very thorough...

RBarryYoung (2/28/2010)

---

Yeah, I was futzing over whether to do that or not, as it can have...

Paul White (2/27/2010)

---

I wouldn't mind one or two more, though. I'm primarily looking for feedback on style and readability now - not technical stuff, as I think that's pretty...

Thank you very much for...

Barry,

Just a quick reply because it is late.

To answer your second question: FLWOR is specified as an XQuery within a query method, so it returns an instance of untyped XML....

Barry,

A very small thing regarding your script: if you are going to issue DROPCLEANBUFFERS, you should always issue CHECKPOINT first. Otherwise, you're just dropping unmodified data from the Buffer...

Looks like a data corruption problem.

Restore from backup.

😀

I don't normally post links to commercial software, but since this site is funded by redgate:

http://www.red-gate.com/products/sql_backup/row_level_restore/why_row_level_restore.htm

Paul