On Apr 20, 8:32 am, Jay wrote:
Hi Spy128Bit -

Good show. Your persistence paid off with the filtering approach.

Interestingly though, I ran your procedure(s) and couldn't get your reported
performance on my test data; no matter, but I'm curious why. It might be the
way I ginned up 32000+ records or some other reason. Below is my version.
It processes 32,400 records in 4 minutes and 15 seconds on my PC (2.66 Core 2
Duo, 2GB Ram, XL2003/WinXPPro-SP2). I expect our platforms would be fairly
similar in performance.

Could you run my procedure on your data and report back with execution time
? It would be of general interest, but also it's an opportunity to shake
down performance concepts.
--
Jay

Our systems are pretty close but the results... jaw dropping. Using
your macro it took 27.1 <bseconds</b. I ran it twice to make sure
and verifeid our results are the same. I don't think WOW can cover it
enough. I will be spending a good bit of time looking over what
you've provided to learn from it. Yeah, WOW is all I have to say
right now. Ok, I need to get some air now. I can't thank you enough
for posting your version to test. I will be instantly switching to
yours now. Thanks a ton! You've been an incredible help.