Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!usc!apple!excelan!crdgw1!brspyr1!tim
From: tim@brspyr1.BRS.Com (Tim Northrup)
Newsgroups: comp.sys.pyramid
Subject: Re: strange 'sort' action
Message-ID: <6541@brspyr1.BRS.Com>
Date: 30 Nov 89 14:32:42 GMT
References: <1155@Terra.cc.brunel.ac.uk>
Organization: BRS Information Technologies, Latham NY
Lines: 31

linda@cc.brunel.ac.uk (Linda Birmingham) writes:
>When we restrict sorting to a particular field it still insists on sorting
>on the first field afterwards.
>Try    sort +1 -1.4 try
>where try is:
>wwww 0020 ctre ge
>tttt 0020 atre ge
>and notice that the tttt line is placed before the wwww line.
>Linda.

This action seems correct to me, at least on my interpretation of the
man page for sort(1).  From the System V User's Manual:

	"Lines that otherwise compare equal are ordered with
	 all bytes significant."

This is after the section discussing sort keys.

In my own opinion, since you only requested a sort on the 2nd word of
each line, you should consider the order of the remainder of the
fields as undefined!  I would expect the sort algorithm to only sort
on the fields I requested it to, and if 2 records are equal for those
tested fields then they would be placed on the output in whichever
order makes the sort program happy.  Just my 2 cents worth ...

						-- Tim
-- 
Tim Northrup      		  +------------------------------------------+
+---------------------------------+         GEnie:  T.Northrup               |
UUCP: uunet!crdgw1!brspyr1!tim    |   Air Warrior:  "Duke"                   |
ARPA: tim@brspyr1.BRS.Com	  +------------------------------------------+