I recently ran into an issue where I needed to diff two files and determine what text was different though I did not want text to be marked as different if it appeared on different lines. As an example, if I had the below two files, I wanted diff to return nothing
1 2 3 4 | $ cat file1 Some text Blah $ cat file2 Blah Some text |
To my surprise, when I ran diff it returned:
1 2 3 4 5 | $ diff file1 file2 1d0 < Some test 2a2 > Some test |
Looking at the diff man page, the only relevant option I could find was –ignore-all-space, but this flag had no impact on the result of the command. The reason why this option did not work was because it only removed spaces on the same line between the files. For example:
1 2 3 4 5 6 7 8 9 10 11 | $ cat file3 Some text $ cat file4 S o me tex t $ diff file3 file4 1c1 < Some text --- > S o me tex t $ diff --ignore-all-space file3 file4 $ |
Now, you may be wondering why I desired the functionality to diff two files with the same text on different lines. I was trying to create a configuration synchronization script, but I learned that the configuration output was different depending on what order commands were executed on the device. In my particular case, the two devices I was comparing were initially configured differently so the lines did not always match up.
So how do you compare two files where the text is the same, but the text appears on different lines?