tr helps you to map one set of characters to another set of characters. Features like range, repeats, character sets, squeeze, complement, etc makes it a must know text processing tool.

To be precise, tr can handle only bytes. Multibyte character processing isn't supported yet.


Here's some examples that map one set of characters to another. As a good practice, always enclose the sets in single quotes to avoid issues due to shell metacharacters.

# 'l' maps to '1', 'e' to '3', 't' to '7' and 's' to '5'
$ echo 'leet speak' | tr 'lets' '1375'
1337 5p3ak

# example with shell metacharacters
$ echo 'apple;banana;cherry' | tr ; :
tr: missing operand
Try 'tr --help' for more information.
$ echo 'apple;banana;cherry' | tr ';' ':'

You can use - between two characters to construct a range (ascending order only).

# uppercase to lowercase
$ echo 'HELLO WORLD' | tr 'A-Z' 'a-z'
hello world

# swap case
$ echo 'Hello World' | tr 'a-zA-Z' 'A-Za-z'

# rot13
$ echo 'Hello World' | tr 'a-zA-Z' 'n-za-mN-ZA-M'
Uryyb Jbeyq
$ echo 'Uryyb Jbeyq' | tr 'a-zA-Z' 'n-za-mN-ZA-M'
Hello World

tr works only on stdin data, so use shell input redirection for file input.

$ tr 'a-z' 'A-Z' <greeting.txt

Different length sets

If the second set is longer, the extra characters are simply ignored. If the first set is longer, the last character of the second set is reused for the missing mappings.

# only abc gets converted to uppercase
$ echo 'apple banana cherry' | tr 'abc' 'A-Z'
Apple BAnAnA Cherry

# c-z will be converted to C
$ echo 'apple banana cherry' | tr 'a-z' 'ABC'

You can use the -t option to truncate the first set so that it matches the length of the second set.

# d-z won't be converted
$ echo 'apple banana cherry' | tr -t 'a-z' 'ABC'
Apple BAnAnA Cherry

You can also use [c*n] notation to repeat a character c by n times. You can specify n in decimal format or octal format (starts with 0). If n is omitted, the character c is repeated as many times as needed to equalize the length of the sets.

# a-e will be translated to A
# f-z will be uppercased
$ echo 'apple banana cherry' | tr 'a-z' '[A*5]F-Z'

# a-c and x-z will be uppercased
# rest of the characters will be translated to -
$ echo 'apple banana cherry' | tr 'a-z' 'ABC[-*]XYZ'
A---- BA-A-A C----Y

Escape sequences and character sets

Certain characters like newline, tab, etc can be represented using escape sequences. You can also specify characters using \NNN octal representation.

# same as: tr '\011' '\072'
$ printf 'apple\tbanana\tcherry\n' | tr '\t' ':'

$ echo 'apple:banana:cherry' | tr ':' '\n'

Certain commonly useful groups of characters like alphabets, digits, punctuation, etc have named character sets that you can use instead of manually creating the sets. Only [:lower:] and [:upper:] can be used by default, others will require -d or -s options.

# same as: tr 'a-z' 'A-Z' <greeting.txt
$ tr '[:lower:]' '[:upper:]' <greeting.txt

To override the special meaning for - and \ characters, you can escape them using the \ character. You can also place the - character at the end of a set to represent it literally. Can you reason out why placing the - character at the start of a set can cause issues?

$ echo '/python-projects/programs' | tr '/-' '\\_'

info See tr manual for more details and a list of all the escape sequences and character sets.

Deleting characters

Use the -d option to specify a set of characters to be deleted.

$ echo '2021-08-12' | tr -d '-'

$ s='"Hi", there! How *are* you? All fine here.'
$ echo "$s" | tr -d '[:punct:]'
Hi there How are you All fine here


The -c option will invert the first set of characters. This is often used in combination with the -d option.

$ s='"Hi", there! How *are* you? All fine here.'

# retain alphabets, whitespaces, period, exclamation and question mark
$ echo "$s" | tr -cd 'a-zA-Z.!?[:space:]'
Hi there! How are you? All fine here.

If you use -c for translation, you can only provide a single character for the second set. In other words, all the characters except those provided by the first set will be mapped to the character specified by the second set.

$ s='"Hi", there! How *are* you? All fine here.'

$ echo "$s" | tr -c 'a-zA-Z.!?[:space:]' '1%'
tr: when translating with complemented character classes,
string2 must map all characters in the domain to one

$ echo "$s" | tr -c 'a-zA-Z.!?[:space:]' '%'
%Hi%% there! How %are% you? All fine here.


The -s option changes consecutive repeated characters to a single copy of that character.

# squeeze lowercase alphabets
$ echo 'hhoowwww aaaaaareeeeee yyouuuu!!' | tr -s 'a-z'
how are you!!

# translate and squeeze
$ echo 'hhoowwww aaaaaareeeeee yyouuuu!!' | tr -s 'a-z' 'A-Z'

# delete and squeeze
$ echo 'hhoowwww aaaaaareeeeee yyouuuu!!' | tr -sd '!' 'a-z'
how are you

# squeeze other than lowercase alphabets
$ echo 'how    are     you!!!!!' | tr -cs 'a-z'
how are you!