NetNews Usenet Archive 1992 #18/NN_1992_18.iso/spool/comp/unix/question/10029

Internet Message Format | 1992-08-12 | 1.7 KB

Path: sparky!uunet!elroy.jpl.nasa.gov!usc!rpi!usenet.coe.montana.edu!news.u.washington.edu!uw-beaver!fluke!ssc-vax!brennan
From: [e-mail] (Mike Brennan)
Newsgroups: comp.unix.questions
Subject: Re: A few UNIX questions (sed+awk)
Message-ID: <[e-mail]>
Date: 12 Aug 92 18:11:27 GMT
References: <[e-mail]> <[e-mail]>
Organization: Boeing Computer Services
Lines: 36


In article <[e-mail]>, [e-mail] (Uwe Waldmann) writes:

| > In article <[e-mail]>
| > [e-mail] (Sam Atabaki) writes:
| > >2. I want to write a little "sed" script (I don't have perl) to
| > >   edit patterns of the following forms:
| > >      xyz abc <[e-mail]> kjkjkj
| > >      xyz abc <[e-mail]>
| > >   such that the <> are removed and anything delimited by <> be
| > >   moved to the beginning of the line. So after the edit the two
| > >   lines above will look:
| > >      [e-mail] xyz abc kjkjkj
| > >      [e-mail] xyz abc
| >

| The following code is more reliable (and it's only one process):
|
|   sed -e 's/^\(.*\)<\(.*\)> *\(.*\)$/\2 \1\3/'
|
| By the way, the sed+awk solution in article <[e-mail]>
| by [e-mail] (Aakash Sahai) is even worse:  It also fails
| if the <...> part is not enclosed by whitespace or if it is followed by
| more than one word.  Why do so many people believe that string substitution
| problems can be solved most easily using awk?

Because it is often easy, for example, this problem with an up to date
version of awk as defined in 1988 AWK book:

awk  'BEGIN{ FS = " *<|> *"} ; { print $2, $1, $3 }'

--
Mike Brennan
[e-mail]