If you're not already cycling your error log, you should.

Daily seems a bit high. We cycle ours once a week.

Boy. I don't know where I've been. I sure thought it was that way. Maybe I'm channeling some old bit of knowledge from a different system. Damn. I hate being...

You have to start from a concurrency model. It's either pessimistic or optimistic. You seem to be assuming an optimistic model, there are unlikely to be very many or any...

That's about as open ended a question as it's possible to ask.

Obviously, the Books Online would be a great place to start. I'd also suggest perusing Itzik Ben Gan's TSQL...

If it's possible, format the delimited list as XML and use XQuery to retreive it from there. We've started to find this a more optimal approach than the old one...

Not wrong, at all, but with something this large, I'd suggest the possibility of looking at exporting to a file and then using BULK INSERT to move across. It's a...

Great comments. That is one of the most difficult problems we encounter working with development teams. They don't trust us and we certainly don't trust them. It takes a while...

A large part of how the optimizer gets the data out of the tables faster is statistics. When you're creating artificial tables, such as table valued functions, there are no...

Ah, I wasn't aware that changed in 2005. Thanks for letting me know. That'll change decisions in the future.

Personally, I kind of don't care, bit or tinyint, but if something...

I'd say it's because it can, and probably did, put statistics on the columns in the temp table that you did with calculations in the other queries. That would make...

It makes sense, but there's not automated way to do a data cleanup like you're asking for. You either have to clean the data manually, or you need to toss...

The main reason I've received for not using bit fields is because they don't allow null values. You have to put a 0 or 1. That's why you may be...

Steve's much more accurate than I am. You have to remember, I spend most of my time working in development, not production. That makes me a lazy b*****d.

What about collation?

If there is a difference between the databases that might cause some odd behavior, reordering columns, affecting how it deals with null values...mabye... It's just a SWAG.

Partitioning is a whole different discussion, that I'm frankly not qualified to take part in.

As to the size of the log... I treat it the same as the data. I...