[SOLVED] Replace a whole string which contains a spefiic word(s) in it

abubakar-malik · 02-14-2011, 04:39 AM

Hi All,

I am quite new in Linux same as with 'Sed' command. I have found most of my solutions on the forums here except this last one. Plz guide me!

I have following input lines in my source file;

07/02/2011/105139 10.0.117.41:58251 3456.52
07/02/2011/105152 P17StreamDistributorI15BlockingLFQueue9AbsThreadE0x2aaad3e4f460 3849.7

I want to replace every string which contains 'StreamDistributor' word in it, with "CTF-Stream--Distributor" so the desire output should be as follows;

07/02/2011/105139 10.0.117.41:58251 3456.52
07/02/2011/105152 CTF-Stream--Distributor 3849.7

I will prefer to use 'Sed' if its possilbe. Any idea????

Regards,
A.Malik

grail · 02-14-2011, 04:54 AM

Would probably depend on if the format is always as you have shown, ie does 'StreamDistributor' always appear in the second column of information delimited by a space?

Also, what have you tried so far?

abubakar-malik · 02-14-2011, 05:29 AM

Thanks for reply grail.
I already have found solution

Here is it ;

Code:

echo 07/02/2011/105152 P17StreamDistributorI15BlockingLFQueue9AbsThreadE0x2aaad3e4f460 3849.7 | sed 's/P17StreamDistributor[^]]* /StreamDistributor /g'

grail · 02-14-2011, 07:46 AM

Here are a couple of alternatives for you to think about:

Code:

sed -r '/StreamDistributor/s/^([^ ]* )[^ ]+( .*)$/\1CTF-Stream--Distributor\2/' file

awk '$2 ~ /StreamDistributor/{$2 = "CTF-Stream--Distributor"}1' file