Christopher Ford (2/13/2008)

---

Earlier Jeff posted a question on using SUM(SAL) OVER()

Since OVER() doesn't support ORDER BY when used with an aggregation.

How about this one?

SELECT J1.AccountID,

J1.Date,

...

I am not sure of any hard and fast rules, though I am sure they exist! Check with sp_configure to make sure you haven't got it limited to a...

could be fine - as long as they don't pop open more than a couple of connections each at any one time.

any idea how many connections your server can sustain...

BOL 2000

"sp_attach_db should only be executed on databases that were previously detached from the database server using an explicit sp_detach_db operation. If more than 16 files must be specified, use...

Try looking at create database for attach.

You cant restore a database with the same name as an existing database - but you can restore it with a different name, apply...

yeah, but we aren't talking about _small_ performance gains here... more like bringing a 6+ hour process that nails the server down to under a minute... with negligable load.

I would...

Pure Awesomeness!

Run over 1.4million records on a 6 level 'hierarchy' in about 10 s - compared to the 'set' solution that takes 1.5hrs...

:exclamationmark:

Oh yes, and validation against the previous...

For what its worth, I remembered a thread about this on google a while back.

I think it was this one, if this link works!

Covers the main points of the argument...

Thanks Hugo for a very concise and informative post, far better than that which I was going to attempt!

It would be nice if the initial article could be...

Not the function I'd worry about - more the update of 4 billion + records in...

quite apart from how succesfull you may with taking down your server trying it...:P

Have you checked for file fragmentation? Sounds obvious, but the number of times I've had to have an argument with someone about whether to defrag a raid drive or...

could be, but how?

Also why would I get a different result running the query to create a job in QA connected from one server, and running it in QA on...

haha - no... I wonder if there's something catching going about tho...?

That could have been an easier solution 😀

Still its nice to get a restart in every so often :hehe: