CODEWITHSUNDEEP
unixsed
Ankur
Ankur

Reputation: 33

Replacing the last word of a line only if match string found

I want to replace the last word of the line only if a matching string found.

Input file :

 "id": 5918915, 

 "description": "Test Job - NA", 

 "revision": 5

Expected output :

 "id": 5918915, 

 "description": "Test Job - EU", 

 "revision": 5

So, for lines matching description, replace the last word with given word. In this case, in line 2 replace last word NA", with EU",

I tried 

sed -i '/"description"/s/.*/EU",//g' file_name

but it is not working

Upvotes: 2

Views: 2006

Answers (2)

Ankur
Ankur

Reputation: 33

I got the working command

sed -i '/"description"/ s/[^ ]* *$/EU",/' file_name

Upvotes: 1

user unknown
user unknown

Reputation: 36229

sed -i -r '/^[ \t]*"description":.*/s/^(.* )NA",[\t ]*$/\1EU/'  FILE

 "id": 5918915, 
 "description": "Test Job - EU",
 "revision": 5"

For testing, remove the -i switch.

The amount of whitespace isn't quiet clear, so I placed [ \t]* at line start and end for blanks and tabs of random size.

Your command:

  sed -i '/"description"/s/.*/EU",//g' file_name

should substitute the whole line with EU",, not just the last char sequence.

The -i is an option of GNU-sed. Check your version and read the fine manual. If your sed lacks support, you have to redirect the output to a file sed "COMMANDS" INFILE > TMPFILE; mv TMPFILE INFILE. Note, that sed "COMMANDS" INFILE > INFILE will not work, but destroy the INFILE immediately; a popular, clever, but disfunctional idea. I had it too. :)

Upvotes: 2

Related Questions