RE: another can this be done with sql

> From: R. Bruce Hoffman, Jr.
>
> or might not have been the right word...
> The third join would look for a record that contains a particular
> combination of stat _and_ acctstat _and_ sotype.
> The files would look for any valid stat and then any valid
> acctstat and then
> any valid sotype.
> It would result in a much more inclusive join.

Cool.  I understand.  The issue here is whether each of the tests is
dependent or independent.  That's not entirely clear from Mike's original
post, because he mentioned having a single file of all combinations.
However, he also mentioned three separate files, which leads me to believe
that the tests are independent.

If the tests are dependent (that is, the valid values for sostat depend on
the value of acctstat), then a single file with valid combinations makes
more sense.  But if the tests are independent, and you use a single file,
then you have to populate the file with every possible combination.  Not
necessarily a bad thing when you just have a few values, but it could get
ugly as the number of values grows and the maintenance of the file could
lend itself to errors - unless you wrote a second program that populated the
"valid status combination file".

Mike, how about a little more information as to exactly how the tests work?
Does the selected record just have to pass the three tests independently?
That is, do you just check one field at a time against a set list of values,
and if it passes all three tests, the record is good?

Joe