"Jason Morin"
wrote in message ...

Well, with the examples you gave, my example worked
pretty well. It's obvious that you data is much more
inconsistent in format than your original post lead me to
believe.


Agreed - I do appreciate your assistance.


What you're looking for is an extremely complex
algorithem to handle such wide disparities in the data.
Even then, you won't get close on some of them. I think
the solution lies upstream in the process, before
receiving the data. If you have any influence on your
sources, you would try to insert some type of consistency
in the way they report their data, so that you don't end
up in the mess you're in now.


Unfortunately the data is already in existence for many periods and
that is what we have to work with.

The systems going forwards are fully integrated so the issue does
exist in the same way.


I'd suggest applying what I've given you, enhance it as
much as you can, then suck it up and do a manual check on
the rest.

Jason


I wish we could afford the manual time, it would be easier from a
management perspective at least - I will keep on looking for an
algorithm, but thank you again for your assistance.

Alan.