UNIX for Dummies Questions & Answers

Hi,

I'm trying to update the last two characters coming in a string globally in a file.
Here is the sample data:

file1
In file1, I want to have all instances replace where _o is appearing in the end of a word with _g. If _o is appearing in the middle or any other position except the end, shouldn't get replaced.

So, the output should be:

The below one liner isn't working as it replaces all instances of the above mentioned string.


How can I accomplish this i.e. replacing the instances where the instance of the string is coming at the end of each word.

Thanks

You are very close. You need to include an 'end of word' marker so that it finds only the _o string at the end of the word.

Code:

sed -r 's/_o([ \t])/_g\1/g; s/_o$/_g/'

assumes that all words are terminated with either a space, tab, or newline. The use of the parens and \1 in the replacement, causes that part of the match to be used in the substitution; thus if a word ends in a tab, the replacement will be _g<tab>. If you have more characters that define the end of the word you can add them to the class with the space and \t.

Replace -r with -E on the command line if you are using a BSD flavour of sed, or are using sed from AT&T's AST distribution.

Code:

perl -wpl -i.bak -e 's/_o$/_g/g' file1

But in your sample O/P, some _o followed with space is also updated.

Code:

 sed -e 's/_o$/_g/;s/_o /_g /' urfile