Solved: odd sort of bit-counter...

["Dan Bale" <dbale@xxxxxxxxxxxxx>]

(cross posted to RPG400-L, midrange-L, and MI400)

For those who were following this thread for the technical aspects, I have a
workable solution for the "subset sum problem" written in RPG.  If anyone is
interested, I would be happy to share the code for peer review.  If I get a
response to this on the RPG400 list, I'll post it there.  Otherwise, I'll
just send it directly to those who request it.  (There's 73 statements in
the source member.)

Thanks to all for some great ideas!

BTW, as far as the business case for this type of solution is concerned,
there is, in fact, no matching detail between the two amount lists.  What
happens is that the A/R people enter a payment into Mapics (DB).  Mapics, by
default, applies the payment to many invoices behind the scenes.  By the
time it is posted, we lose the fact that a total payment of x dollars was
made; we have only all of these smaller amounts applied to different
invoices.  The A/R people don't normally concern themselves with the
invoices that the payments are applied to, so that is why I'll get a list
that I can't duplicate verbatim from the information stored in Mapics.

BTW, this is not an unusual problem with Mapics DB.  You should see the
hoops I have to go through to print a month-end summary of the A/R Invoice
Register.

db