Just keep in mind that data compression (which is essentially what you're looking at there) only goes so far, while multiple queries can scale infinitely to as many tables and...

Think of it as "you hit them first after they kicked you".

You're welcome.

I prefer not to use mixed mode, because that means the sa login is active. Find that one password, and you own the server. Windows authentication, you have...

In my experience, OpenRowset is the easiest way to do that.

What issues are you having with that?

I always use the From version. Allows for more complex logic, since In only allows comparison of one column at a time unless you get into correlation complexities.

I alias...

First, datetime isn't quite that accurate. You might need to make the milliseconds 997 to get what you're looking for.

Second, depending on how the data gets into the field,...

Any reason to not use more than one OpenQuery and just join them together?

Those seem like legitimate questions to me.

If the sequence matters, put it in the trigger code, and don't count on firing sequence.

Most of the time, it's better to have the code in the insert proc instead...

The only way you'll get an index on every column is to vertically partition the table (move some of the columns to separate tables), or build a lot of indexed...

I don't know. I see far too many statistical analyses of various groups of people, and I just don't see the use of them in the vast majority of...

Yes, there are a lot of ways to accomplish this. I still think the simplest is to use "not" for the test.

It sounds like what you're looking for is parsing the data, then pivoting it.

The data from this article will teach you how to parse it http://www.sqlservercentral.com/articles/T-SQL/62867/.

That will give you one...

The first thing I'd do with this would be to switch to ANSI style joins, where you explicitly list the join type and the join math in the From clause,...