Conspiracy? Frankly I stay away from sorts these days. We run in the 36 environment, and there are quite a few sorts on those old OCL's. When I add stuff and, I use LF's or SQL. I have a hard copy of the Sort manual, which we used, when I was working on a Y2K package, to figure out how to change FMTDTA and $GSort specs. But I haven't referenced it since.
I do remember back in the S/3 days writing some sorts that were small programs in their own right. There and on the SA/34/36 we often had to use what are called "record address" files (basically RRN's into the PF). These did require some special coding in both the F- and E-specs. We used those because often there wasn't enough disk space for a separate output file, much less the work file. I vaguely remember tricks one could use, such as defining the work file location in the OCL (usually programmers just let it default to wherever IBM thought best).
One could, also, do a lot of complicated selects/omits and within each set use 'F'orce specs, alternate sequence tables, and on and on. I've seen (and written, I must admit) sorts that would drive a man to drink (assuming that one needed a reason, of course).
Jerry C. Adams
IBM System i Programmer/Analyst
From: midrange-l-bounces@xxxxxxxxxxxx [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Vern Hamberg
Sent: Tuesday, September 01, 2009 5:59 PM
To: Midrange Systems Technical Discussion
Subject: Re: AS/400 FMTDTA (sort) manual
Thanks Steve - I seem to remember a recent thread about some S/36 stuff
having more flags - maybe it was something in one of the RPG spec lines
up near F-specs - Jerry Adams mighta been part of the conspiracy.
It has been a while, but from what I recall there was little or no
difference between the way Sort specs are coded on the S/36 - S/38 -
AS/400 - iSeries - System i - IBM i.
Now that is the AS/400 Sort manual -
I suspect you want some of the old
S/36 stuff, and that is not easy to find.