assuming you have some mailfunctionality properly set up and working (xp_sendmail, xp_smtpmail or other), this is nothing extraordinary at all.

Just loop through your data, for each mailaddress, gather the appropriate...

Can you explain a bit about the underlying problem..? I mean, are you trying to find out how to insert a bunch of rows into a table with existing data,...

If you run the statement from QA, you'll get the offending value(s) back with the results

/Kenneth

heh not going there. I'm sure you're right about that.

But, since it's x-mas, I'll just mess you all up with my favorite version

DECLARE...

umm.. not to be picky, but it's not stored as a float, it's stored as two two-byte ints..

/Kenneth

You can use replace() on the idcolumns to replace spaces with empty strings.

replace(id1, ' ','') = replace(id2,' ', '')

Note though that this will invalidate any indexusage on those columns enclosed...

From BOL (your best buddy)

BCP

-r row_term

Specifies the row terminator. The default is \n (newline character). Use this parameter to override the default row...

Well, actually what I'm suggesting is to stick to the things as they are - the light load doesn't seem to justify a remake of the design, I believe.

/Kenneth

Not of much help here, I'm sorry to say.. But for the most parts when bcp complains about unexpected EOF, it has to do with the rowterminator not matching what...

imo the most scalable 'arbitrary-multi-row-argument' method is to update (or delete) by joining the base table against another table holding the keys for the rows you want to update or...

Yes, those are single-value statements, since the variables can only contain one value, each delete or update handles only one SARG at a time. (in theory it could be more...

I don't think this is a 'calling issue', but rather a procdesign issue...

How are your procs written? Can they handle multiple-row deletes or are they designed for singleton deletes?...

If actual writing activity in general is wanted, I believe that you could use perfmon for this.. (gotta be some file i/o counters in there, I believe..?)

/Kenneth

Oh, yes, you're absolutely right. The step beyond horrible is ever so short.

I too assumed that userId was unique, hence the note about userId + job must then become the...

awww.. not a concatenated list... that will be absolutely horrible to search through...

I'd go for the userid + job instead, one row for...