Home / ubuntu / Questions / 857284
Accepted
mcExchange
Asked: 2016-12-06 05:43:03 +0800 CST

Replace second instance of string in a line in an ASCII file using Bash

  • 772

I have an ASCII file with the following structure:

file1.png otherfile1.png
file2.png otherfile2.png
file3.png otherfile3.png
...

I want to replace .png with .mat, but only for the second column. The result should be like this:

file1.png otherfile1.mat
file2.png otherfile2.mat
file3.png otherfile3.mat
...

How do I do that in Bash?

command-line

2 Answers

  1. Best Answer

    Well, if it is the end of the line...

    $ sed 's/\.png$/.mat/' file
file1.png otherfile1.mat
file2.png otherfile2.mat
file3.png otherfile3.mat
    • s/old/new/ search and replace
    • \. literal dot (without the escape it matches any character)
    • $ end of line

    Or to explicitly specify the second column, you could use an awk way...

    $ awk 'gsub(".png", ".mat", $2)' file
file1.png otherfile1.mat
file2.png otherfile2.mat
file3.png otherfile3.mat
    • gsub(old, new, where) search and replace
    • $2 second column
    • 13

  2. You can replace all .png strings directly at the end of a line in INPUTFILE like this:

    sed 's/\.png$/.mat/' INPUTFILE

    The command above will not modify INPUTFILE but only print the changed version to the terminal.

    To directly edit the file in place, add the -i flag to sed (or -i.bak to store a backup of the original file):

    sed -i 's/\.png$/.mat/' INPUTFILE
    • 7