[SOLVED] replace string in file using perl or sed

wf201626 · 06-30-2011, 10:17 AM

hi all

I have question about replacing a string in a file.

/etc/hosts

10.1.1.1 printera
10.1.1.1. printerb

how can I replace the printerb's 10.1.1.1 to something else(10.1.1.2, for example) without replace printera's 10.1.1.1 accidentally?

I have tried perl -e -pi "s/10.1.1.1/10.1.1.2/g" /etc/hosts. but, perl replace both 10.1.1.1 to 10.1.1.2.

any idea will be appreciated.

thanks in advance,

F.

colucix · 06-30-2011, 10:35 AM

Code:

sed -i '/printera/s/10.1.1.1/10.1.1.2/' file

or

Code:

sed -i /printerb/!s/10.1.1.1/10.1.1.2/' file

wf201626 · 06-30-2011, 11:02 AM

hi colucix

the sed works for this case. but I want to replace the specific string and apply the change into the file. that is why i used perl.

thanks,

wf201626 · 06-30-2011, 11:09 AM

I found someone used this -

perl -pi -e 's/ReplaceMe/REPLACED/ if /MatchText/' filename

but it does not work for me.

idea?

David the H. · 06-30-2011, 11:11 AM

You should take bit of time to read up on sed. It can do a whole lot more than just simple substitutions.

Here are a few useful sed references.
http://www.grymoire.com/Unix/Sed.html
http://sed.sourceforge.net/sedfaq.html
http://sed.sourceforge.net/sed1line.txt

Read up on awk too, while you're at it, which is often the better tool to use when working with field-style input. It can be used here as well, although you have to move it through a temp file to save the changes.

Code:

awk '/printerb/ { $1="10.1.1.2" }1' /etc/hosts >tempfile
mv tempfile /etc/hosts
rm tempfile

A few useful awk references.
http://www.grymoire.com/Unix/Awk.html
http://www.gnu.org/software/gawk/man...ode/index.html
http://www.informatik.uni-hamburg.de.../gawk_toc.html
http://www.pement.org/awk/awk1line.txt
http://www.catonmat.net/blog/awk-one...ined-part-one/

grail · 06-30-2011, 11:29 AM

Quote:

the sed works for this case. but I want to replace the specific string and apply the change into the file. that is why i used perl

I am not sure you paid attention to the example because the -i option does make the change in the file.

wf201626 · 06-30-2011, 12:14 PM

Quote:

Originally Posted by grail

I am not sure you paid attention to the example because the -i option does make the change in the file.

my sed's version has no -i option at all. it is aix.

David the H. · 06-30-2011, 12:23 PM

Well, don't you think that's something you should've mentioned at the beginning, then? You're probably not using bash either, right?

It's generally a good idea to spell out the full context of your situation, so we can provide you with useful answers, without wasting effort on things you can't use.

In any case, what's wrong with working through a temp file, as in my awk example above? It's only a couple of extra lines.

Oh yeah, and saying "it does not work for me", in your perl question above isn't very useful either. HOW did it fail? What exact command(s) did you use, and what output or error messages did you get?

wf201626 · 06-30-2011, 12:26 PM

sorry about that, David. we are using ksh in AIX environment.

let me think about your way.

wf201626 · 06-30-2011, 02:43 PM

here is what i got -

10.1.1.1 printera
10.1.1.1. printerb

perl -piw -e 's/10.1.1.1/10.1.1.2/ if /\bprinta\b/' /etc/hosts. this will find the line contained exact matched key word and replace it the whatever you want to.

this perl one liner in my script will be generated by provided variables.

here is my soluton

#!/usr/bin/ksh

PRIP=10.1.1.1
OPRIP=10.1.1.2
PNAME printera

sub="s/${PRIP}/${OPRIP}/ if /\b${PNAME} \b/"
echo "perl -piw -e " "'${sub}'" " /etc/hosts" | ksh

grail · 07-01-2011, 02:05 AM

Firstly, please use [code][/code] tags to help make your code readable.

May I ask the point of the following line:

Code:

echo "perl -piw -e " "'${sub}'" " /etc/hosts" | ksh

Why not just call your command?

Code:

perl -piw -e "${sub}" /etc/hosts

Also if you wish to try David's awk without a temp file you could do:

Code:

#!/usr/bin/ksh

PRIP=10.1.1.1
OPRIP=10.1.1.2
PNAME=printera

awk -voprip="$OPRIP" -vpname="$PNAME" '$NF == pname{$1 = oprip}{print > FILENAME}' /etc/hosts

colucix · 07-01-2011, 02:22 AM

An alternative is the ancient ed editor:

Code:

#!/bin/ksh
ed -s testfile <<-EOF
   /printerb/s/10.1.1.1/10.1.1.2/
   w
   q
EOF

wf201626 · 07-04-2011, 08:59 AM

Quote:

Originally Posted by grail

Firstly, please use [code][/code] tags to help make your code readable.

May I ask the point of the following line:

Code:

echo "perl -piw -e " "'${sub}'" " /etc/hosts" | ksh

Why not just call your command?

Code:

perl -piw -e "${sub}" /etc/hosts

Also if you wish to try David's awk without a temp file you could do:

Code:

#!/usr/bin/ksh

PRIP=10.1.1.1
OPRIP=10.1.1.2
PNAME=printera

awk -voprip="$OPRIP" -vpname="$PNAME" '$NF == pname{$1 = oprip}{print > FILENAME}' /etc/hosts

hi Grail

this one liner is ok to run in command-line directly. if you run it in a script,

Code:

perl -piw -e "${sub}" /etc/hosts

will not run correctly.

I will give awk a try and let you know how it works in my env.

thank, Grail

grail · 07-05-2011, 02:29 AM

Quote:

will not run correctly.

Well I do not have access to ksh at the moment, but it should not matter as you essentially are not using
anything ksh-centric, and it works just fine inside the script (as I would expect seeing the script is just a wrapper).

When you say it doesn't run correctly, what do you mean?

Here is what I tested on:

infile:

Code:

10.1.1.1 printera
10.1.1.1 printerb

script.sh:

Code:

#!/bin/bash

PRIP=10.1.1.1
OPRIP=10.1.1.2
PNAME printera

sub="s/${PRIP}/${OPRIP}/ if /\b${PNAME}\b/"

perl -piw -e "${sub}" infile

And when run:

Code:

$ ./script.sh
$ cat infile
10.1.1.2 printera
10.1.1.1 printerb