Re: Knotty puzzle with chaining to files not known at compile time

[Booth Martin <booth@xxxxxxxxxxxx>]

To add two cents at this late date:   Is this a place where an 
intermediate file makes sense?  Define the intermediate file with a key 
field that is as long as the longest key, and only the data fields you 
actually will use, including a redefinition of the fields that are used 
as keys if you need them for anything other than ordering the records.

Then, you have two problems left to deal with.  The simple one of using 
the data, and the other one of populating the new intermediate file, 
which could be done I suppose with RPG, SQL, or #GSORT?



James H H Lampert wrote:
> Clarification:
>
> There could be as few as zero of these files, or as many 
> as 25. The variations in record formats are virtually 
> limitless, but there are exactly three different 
> possibilities for the first few fields, including both 
> keys, as outlined in the original query.
>
> I've got enough suggestions to where I can construct a few 
> experiments to test the two most viable possibilities; 
> I'll get back to the List with my findings.
>
> --
> JHHL