RE: Cycle is broken (V4R5)

At 05:38 PM 7/17/02, you wrote:
>I don't have a manual that old, but I don't think you could EVER use level
>breaks on binary data.  Going back to System/3.  The V5R1 manual documents
>this on page 181, numeric data type.  I looked a bit farther down and
>discovered that integers cannot be control fields either.  Page 367 of the
>S/36 compatible RPG II (June 1994) manual says that binary cannot be control
>fields.

Yes, I think you're right about the fact that the old stuff wouldn't allow
a binary field to be a control field, but now?  A binary field can be
treated as an alphanumeric for the purposes of a control field.  I just
don't see why it's restricted. IMHO, it should not be.  It was probably
more work for the compiler team to add the check and the error message than
to generate the code for a binary control field.


>If it were me, I'd ask the guy who designed the file how to solve the
>problem.  Sorry, too sour.

As I indicated, it can't be changed, mainly for political reasons.

>How about trying OPNQRYF or a LF that makes it
>zoned or packed?  In the old days I would have used sort (FMTDTA).

Aha!  This is the bright idea I was looking for.  I was able to put
together an LF without much trouble.  Thanks to you and also to Phil who
made the same excellent suggestion.

My last resort was to create a near identical physical and cpyf *map into
it, but this is much better for my situation.



Regards,
Rich