If this is an ad hoc need, I'm thinking a three way inner self join on derived tables (one for each of the conditions) might do the trick a bit...

Since "not working" can mean lots of things, I'll just refer you to this, specifically the sqlwb /setup recommendation, if that seems to be the problem you are having.

It...

Since you have a valid product key, I'd personally just call MS and tell them why you need to do it, and see if they'll assist you. Since this could...

Just out of curiousity, assuming you have no way to avoid doing this, is there a reason you can't do an insert into the sequence table and store scope_identity into a...

And you're going to have a real hard time finding anyone in this thread who stated that they shouldn't be separated if that's what they are. The beef was with...

For a variety of reasons, we've always stored them as varchars, but mostly because I've always stored numbers that weren't used for calculations that way. By the same token, I store phone numbers,...

If you wish to avoid dynamic SQL, and if the columns that you want them to potentially sort on are fixed, then you can also do it with a CASE...

Lynn, it passed the million random values test just fine, as well as performing Sergiy's ".0000 to .1999 step .0001" test flawlessly, just like David J's function. It's looking good...

Initial spot checks look good. It properly handles the negative numbers that I've thrown at it so far. In fact, I like it so much that I've stopped working on...

"This leads me to think that you believe that Round() has some special intelligence that takes into account the above concepts and which makes it work perfectly. Do you know...

Oh, so Round() sees .045000?

Cool. Lynn, your function works great on Sergiy's test then, since it rounds .045000 exactly as expected, based on the rules used for Banker's Rounding. Still...

I didn't ask what it represented, I asked what Round() receives as a parameter. It receives a single number, accurate or otherwise. What is that number based on your test?

If what...

No, Round() gets a single value as its numeric expression, not some nebulous range between .0450000 and .0450009. Hell, if what you claimed was true, Round wouldn't even know what...

Then tell me what the built-in Round() function would receive from the test setup you used for Lynn's BRound function. You can limit it to say 15 digits to the...

You can't even stay on topic.

Does or does not the built-in Round() function receive the exact numeric .045 (with nothing but zeroes following), just as Lynn's function does, versus the...