the reason not to do what Kit suggested is that there are multiple
records/provider as there is a new contract at least once per year
(of course, I suppose that I could "fix up" the provider file and
then do this ....)
No, you would not want to expand the big dataset this way. But you
could do the following. Split up the provider file into those with
contracts ending in 1999, those ending in 2000, those ending in 2001,
etc. -- separate files. Merge them sequentially with the master file,
and update the 'status' variable iff it is missing. In the absence of
some SQL logic, I think that would work pretty well (and would not
require expansion of the master file).
