once again we have a good question marred by that little detail - the world is NOT the same as US locale/format! Please Please, pleeeease get your locale/dateformat thought through!...

Depending on your network setup, have you tried pinging?

number of columns = expected number of commas + 1

something like a while loop would allow you to iterate with charindex to a point after the last #expected#, then a...

if you know the number normally expected, do a charindex, loop up to that number, and if the next one comes back with a value, you've got the position.

test for the extra comma, replace it with something else, replace the remaining commas, return the remaining one to a comma.

If you could get them to build the csv with another delimiter - ; or | forinstance, that stops the comma casuign a problem. You'll have to change the dts...

so, tell me. those error codes...

Truly the kind of outfit you #don't# want to work for.

I can see Dilbert written all over this...

SELECT * FROM tblUserAccessLevel l INNER JOIN tblUserContentArea c ON ((l.UserID = @user-id) AND (c.UserID = @user-id))

SELECT @testaccess = Userid FROM tbluseraccesslevel where Userid =

and the error codes?

hum. my PERL is not too strong, but are you confirming what is actually being sent to the proc? as you've got trans/commit lines in there, it might actually be...

ah! you're looking for the userid value?

Its trying to give you back two sets of data at the end

#####

SELECT * FROM tblUserAccessLevel l INNER JOIN tblUserContentArea c ON ((l.UserID =...

but running it through query analyser returns a value?

just a thought, but have you tried to check what data is actually being returned? Like a NULL, or something else that the script doesn't like?