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

Without going into a lot of research, it seems to me (and I could be dead
wrong on this) that a program-described keyed file doesn't care much about
the keys other than the total length of the key.  In fact, in your case, I
would hazard a guess that you simply need to define files with key lengths
of 14, 16 and 18 (and with some suitable record length).

Then you might be able to override to the appropriate file dynamically and
then simply do a SETLL to determine the existence of a given key (which you
would format as a single character field).

I don't know this for certain.  I'm really shooting from the hip here.  But
if all of those assumptions held, I'd just have a program that accepts the
file identification information, an 18-character key field and a flag that
says whether the data is 14, 16, or 18 characters.  Then I'd override, open
and SETLL.

Joe

From: James H H Lampert

I've got a knotty little puzzle here:

I have a number of files. We know what their names are at
compile time, but not necessarily which ones actually
exist in any particular installation. We know that the
files are keyed on two fields; the leading key is either
9, 11, or 13 bytes zoned decimal, and the secondary key is
always 5 bytes, but we have no way of knowing at compile
time what the length of the leading key is.

All of this information can, however, be determined at
runtime.

We also don't know much about the record structure at
compile time, beyond the keys.

All we need to know is whether or not records exist for
each of a series of specified composite key values.