Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ukma!gatech!ncar!tank!mimsy!chris
From: chris@mimsy.UUCP (Chris Torek)
Newsgroups: comp.unix.questions
Subject: Re: sort question
Message-ID: <17411@mimsy.UUCP>
Date: 10 May 89 09:02:21 GMT
References: <199448@hrc.UUCP> <810050@hpsemc.HP.COM>
Organization: U of Maryland, Dept. of Computer Science, Coll. Pk., MD 20742
Lines: 51

In article <810050@hpsemc.HP.COM> gph@hpsemc.HP.COM (Paul Houtz) writes:
>Right.  There is no way to do a true column sort using this utility as you
>can on IBM or MPE systems and here is why:   Sort requires a FIELD DELIMITER
>character.   That means that there is SOME character that will never be 
>sorted.

But (as you yourself point out) you can set the field delimiter to
newline, effectively making it vanish, then use the 0.n format to
specify column n.

>Here is the worst one I have seen on Unix.  I converted this myself from a
>sort done on an IBM System/34.  This is a good example of a COMMON type of
>sort done in the commercial world which you never see on Unix:
>
>sort -dt'\012' +0.6 -0.8 +0.13 -0.15 +0.15r -0.17r +0.8 -0.13 ...
>
>This guy sorts the summary file using the newline character as a field 
>delimiter (i.e., no fields), and you can tell what column ranges are 
>being sorted by subtracting 1 from the 'x' field of the 0.x parms.  

Or simply think of columns as numbered from 0 (if you count from 0
to 1023 on your fingers, as I do :-) ).

>If anyone ever tries to tell you that UNIX is user friendly, you can now
>barf on them.  

Why?

(Actually, Unix is not *meant* to be `user friendly'---if that means
`taking occasional users by the hand and leading them from each little
stepping stone on to the next'.  It is meant to get the job done
simply, tersely, without back-talk.  If you use a system every day,
you can get tired of wading through six levels of menus.  And if Unix
looks a little old and creaky in the user-interface area, well, it
*was* designed around printing terminals and dumb CRTs.  But then
again, that is all I have at home.  [An H19 with the Heath ROMs can
hardly be called clever :-) .])

If this sort of thing is not done often in Unix, why not?  Perhaps
because it is a bad idea not to delimit fields.  (I believe that any
fixed limit---this includes fixed field widths---is always too small.
I sometimes wonder what IBMers do about people with last names like
`de Martinesquez y de la Capillostraglio'; I *know* what they do with
people who, like me, put down an initial for a first name and a name
for a middle initial.)  But if, like Houtz, you are forced to make the
best of a bad design, and you dislike all the `+0.x -0.y', instead of
sulking, you *could* write a small shell script to convert whatever
column format you prefer into what sort requires.
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7163)
Domain:	chris@mimsy.umd.edu	Path:	uunet!mimsy!chris