tr - A Utility for Replacing Characters
This is a simple and very useful utility. Basically, tr filters strings one character at a time, and when used correctly it is even UTF-8 compatible.
Character filtering is defined by the parameters, and to make things easier you can use character classes, for example [:upper:] for all uppercase letters or [:cntrl:] for all control characters.
Parameters
The parameters I use most are the following:
-ddelete to remove characters.-ssqueeze, which is quite handy for collapsing sequences of repeated characters.
Examples
Converts uppercase to lowercase, but fails with UTF-8 characters.
echo ÁÉÍÓÚ | tr 'A-Z' 'a-z'
Same as the previous example, converting from uppercase to lowercase, but this time it doesn’t fail with UTF-8 characters.
echo ÁÉÍÓÚ | tr '[:upper:]' '[:lower:]'
Replaces line breaks with spaces.
echo "teste1\nteste2" | tr '\n' ' '
Strips everything from the string, leaving only the digits.
echo "O numero de telefone é (11) 12345-6789" | tr -cd '[:digit:]'
Removes all control characters, leaving only the plain text.
tr -d '[:cntrl:]'
Takes the contents of the PATH environment variable and breaks the paths one per line.
echo $PATH | tr ':' '\n'
Removes repeated characters.
echo teste teste | tr -s ''
Rot13, probably the weakest encryption in the universe, but still fun.
tr a-z n-za-m
Portability
tr has been rewritten many times for different platforms, and some parameters and behaviors that exist in certain versions don’t exist in others. If you’re concerned about the portability of your scripts, try to stick to the parameters defined by POSIX.
Origins
I believe the first release of the tr utility was in the 1973 version of Unix Time-Sharing System 4.
Authors
- Douglas McIlroy (doug)
References
- UNIX manpages
- A Research UNIX Reader: Annotated Excerpts from the Programmer’s Manual, 1971-1986