Yep... I'd test it... "it depends" and the age old adage of "one measurement is worth a thousand speculations" still holds true.  It's especially true in the world of multi-processors......

The following is wrong because it leaves out one entire second of the day...

BETWEEN '2006-05-01 00:00:00.000' AND '2006-05-01 23:59:59.000'

...depending on the source of the data, there could be thousands...

The reason I ask is because I DO have a Clustered Primary Key on the N column of my Tally table and here's the results I got from the 3...

Kenneth,

Does your Tally table have a Primary Key?

Jeff,

Great use of a Tally/Numbers table... is more than twice as fast as the WHILE LOOP method...

Shifting gears.... in your code, change this...

select @address = '1300 teseter 234'

...to this...

select @address...

Nope, no experience with GPS phone providers here, either.  But what you have planned  not only sounds very feasible, it sounds like a hell of a lot of fun, too! ...

Here's the example xp_CmdShell code I will sometimes use... note that I always include the hardcoded DOS command DIR and I always include the path in double quotes to help...

Not sure exactly how you would do it in DTS but I'm thinking that if you built some error logic around the fact that the files are empty and that...

Another country hear from!  Also, very neat, David.  Interesting use of SIGN as a "lump detector".

Juliano... has one of these 3 solutions provided you with the answer you needed?  They...

... except to continue Shows how much I use loops and I didn't read the code well enough..  Thanks for the clarification...

VERY cool... I've seen a lot of hierarchical code in my time... this is great...  First time I've seen anyone make good use of the CONTINUE statement without them getting...

Heh... I sure do appreciate what that's all about... been there and done that... just not with a cursor so I have no idea if a cursor can be created...

Why do you need a cursor for this?  If you converted the cursor to set based code, the dynamic SQL would work just fine...

Not a bad function but it takes more than 4 minutes to execute on a mere million rows where each row contains a six word value...

The following function, butt ugly...

As Ryan pointed out... had a little bug in my code and repaired it.  Managed to condense it a bit, too.  Still takes a very pigish 16 mintues to run...