Replace a specific word every three occurences

udiubu · 12-07-2011, 08:34 AM

Dear all,

I got this problem today.

I have a log file with several lines and columns.
In a column I have for example this succession of strings:

Picture
word3
word3
word3
Picuture
word3
Picture
word3
word3
Picture
word3
word3
Picture
word3

What I need is a command that replace the string "word3" with let's say " bob" only every three occurences of the "word3" itself, as below:

Picture
BOB
word3
word3
Picuture
BOB
Picture
word3
word3
Picture
BOB
word3
Picture
word3

I'm using this command that of course changes all of them, recoursively:

awk '{for (i=1; i <= NF; i++) if ($i == "word3") $i = "BOB"; print $0}' file.log > file1.txt

Any suggestions would be highly appreciated!

All the best,

Udiubu

udiubu · 12-07-2011, 09:12 AM

First solution:

(1) sub every three..

awk '/word3/{c++;if(c==3){sub("word3","BOB");c=0}}1' file.log >file.txt

(2) sub every two..

awk '/word3/{c++;if(c==2){sub("word3","JACK");c=0}}1' file.txt >file1.txt

Cedrik · 12-07-2011, 09:24 AM

Quote:

Originally Posted by udiubu

First solution:

(1) sub every three..

awk '/word3/{c++;if(c==3){sub("word3","BOB");c=0}}1' file.log >file.txt

(2) sub every two..

awk '/word3/{c++;if(c==2){sub("word3","JACK");c=0}}1' file.txt >file1.txt

Nice, but why not do it in one pass, like:

Code:

awk '/word3/{if(!c++){sub("word3","BOB");}if(c==3)c=0;}1' file.log > file1.txt

udiubu · 12-07-2011, 09:41 AM

Hi Cedrik,

That works, too!!
Thanks..

I'm doing these three steps now as I found useful to have these three occurences named with three different names.

Have a nice day.

All the best,

Udiubu

Nominal Animal · 12-07-2011, 10:38 AM

Quote:

Originally Posted by Cedrik

Code:

awk '/word3/{if(!c++){sub("word3","BOB");}if(c==3)c=0;}1' file.log > file1.txt

You could make it even simpler:

Code:

awk '/word3/ { if (++c == 3) { c=0; sub(/word3/, "BOB") } } 1' file.log > file1.txt

Cedrik · 12-07-2011, 10:58 AM

Quote:

Originally Posted by Nominal Animal

You could make it even simpler:

Code:

awk '/word3/ { if (++c == 3) { c=0; sub(/word3/, "BOB") } } 1' file.log > file1.txt

Simpler indeed, nice