ADSM CLI output usually comes with two main formats (a third one,
the output format of the unsupported show command, is considered
of no real importance): table format and stanza format, where the
latter may be forced using the f=d option (if supported by the
specific command). And there are many things annoying, senseless
or even wrong with all of them - esp. when ADSM output is needed
for further processing.

With both formats information entries might be hyphenated and there
are at least three kinds of splits I know of one has to deal with
(personally I use a perl script to accomplish for this).

In one case, where hyphenation of long information entries takes place
at (single or more) white spaces, output is even worthless as part of
the information is destroyed due to the fact, that ADSM CLI output
(esp. table format) is "formatted to death" using white spaces also.

IBM, are you AWARE of this? You should, as I already told you:
Consider a path name as created on PC platforms. This path name may
contain all kinds of characters - including white spaces. If ADSM CLI
decides to use this/these white space/s to hyphenate, then information
about these white spaces is lost (you don't know, whether blanks or tabs
have been removed, and how many of them). Therefore the original path
name cannot be reconstructed in total from the hyphenated path name.

Even if one (ie. me) is spending hours on writing scripts to undo the
mess created by ADSM CLI output hyphenation. Just consider the situation:
When using CLIs for batch processing lines may be as long as needed and
in this case there is absolutely no need for such kind of treatment.

That's why I would suggest a third main format - let's call it record or
raw format - where field and record separators may be set arbitrarily
(defaulting to tab and newline, resp.). And where no data manipulation
(like damned hyphenation) takes place at all. Only such a record format
is suitable for reliable automatic processing.

BTW: ADSM CLI commands should keep to the common standard as for the exit
value, standard output and standard error. Things became better since
those early days when "dsmc q opt" wrote to StdErr and dsmc return values
seemed to be totally random. But even today dsmc and dsmadmc error output
is not sent to standard error as expected, but to standard output instead.

Rainer