[SOLVED] how replace exact word using sed

santhosh.ananthula · 01-27-2016, 07:44 AM

I am having a requirement to replace a word xyz with zxy in a file containing a.xyz and xyz

when i am using the below command its replacing in both the places which is not the exact requirement

sed 's/xyz/zxy/g' names.txt

also tried
sed 's/ xyz/ zxy/g' names.txt it works but it seems to be no exact solution.

Please advice

Thanks

pan64 · 01-27-2016, 08:16 AM

use sed -r and check for word boundary

ondoho · 01-27-2016, 12:44 PM

it's one of those typical stackoverflow questions.
i don't understand why people don't search before asking... if this goes on, the internet will become totally unusable soon.

TB0ne · 01-27-2016, 02:39 PM

Quote:

Originally Posted by ondoho

it's one of those typical stackoverflow questions.
i don't understand why people don't search before asking... if this goes on, the internet will become totally unusable soon.

Yes...I believe it's been brought up here several times. For this very reason.

I have yet to understand why folks can't apply some effort of their own, rather than looking for someone to give them the exact command...it would never occur to me *NOT* to try to solve the problem on my own. Even more entertaining, because this site is typically found through a Google search...which could also have provided an answer.

rknichols · 01-27-2016, 05:10 PM

This particular case is a bit different since "." is a non-word character, so the string "a.xyz" would still satisfy a full-word match for "xyz". Writing a regex with a different notion of word boundary that still covers the case where the "word" is at the beginning or end of the line is not trivial, nor is concocting a Google search for an answer to that problem.