SED Cheat Sheet

Sed is a stream editor on UNIX-like operating systems that is used to filter and transform text.

Sed can be used for both editing files and standard output of programs/operations to stdout.

Standard syntax:

sed [options] commands [filename]

The following are examples of using sed in various situations.

Word replacement (root on Admin):

sed 's/root/Admin/' filename

As a result, the contents of the file will be printed to stdout and root will be replaced by Admin. Unfortunately, the replacement will be made only for the first occurance of root. To have all output processed, use g in the expression.

sed 's/root/Admin/g' filename

Typically, a sed command consists of two parts, separated by the / character. On the left side, the value or expression that interests us is indicated, on the right side, the new value for the left side.

To perform multiple operations use the -e switch:

sed -e 's/root/Admin/g' -e 's/bash/sh/g' filename

Replace the word only in lines containing the desired one (replace root with Admin in lines with the word user):

sed '/user/s/root/Admin/g' filename

To edit a file use the -i switch:

sed -i '/user/s/root/Admin/g' filename

Removing the 5th line

sed -i '5d' filename

To backup the original file use -i.bak

sed -i.bak '5d' filename

Remove the line containing the word games:

sed '/games/d' filename

Remove all blank lines:

sed '/^$/d' filename

Row range display (rows from one to five)

sed -n '1,5p' filename

The file name is optional. With sed you can change the standard output.

Compare standard output:

df -kh

and its modified version:

df -kh |sed 's/%/000000/g'

Using special characters:

You can use special characters:

  • ^ - start of line
  • $ - end of line
  • . - designation of one character

The complete set is not limited to just these three. Sed understands regular expressions very well.

The following command will replace root with Admin on lines that start with user:

sed '/^user/s/root/Admin/g' filename

The following command will replace root with Admin on lines that end with data:

sed '/data$/s/root/Admin/g' filename

In case you need to use any special character in sed ( $, ^, /, space, dot, ', etc.) it must be escaped with \ .

The following construct will replace the word root followed by a space with Admin followed by a colon:

sed 's/root\ /Admin:/g' filename

In the case of using the symbol | (pipe) special characters do not need to be escaped as a separator, but regular expressions will not work in this case

sed 's|root |Admin:|g' filename

Regular Expressions

Remove all digits from output:

sed 's/\[0-9\]\[0-9\]//g'

Duplicate Clipped Value:

sed 's/root/& &/g'

Try executing:

echo '123abc' | sed 's/\[0-9\]\[0-9\]*/&-&/'

The regular expression [0-9]* matches 0 or more digits. The regular expression \[0-9\]\[0-9\]* matches 1 or more digits.

If you need to use the first word from a line of text, mark it with an escaped digit:

sed 's/\([a-z]\*\).\*/\1/'

Check this out:

echo 'abcd qwer zxc 123'| sed 's/\([a-z]\*\).\*/\1/'

In order to swap the first and second word - use the following construction:

sed 's/\([a-z]\*\) \([a-z]\*\)/\2 \1/'

The space on the left side of the expression can be replaced by any other delimiter.

For test:

echo 'abcd qwer zxc 123'| sed -r 's/([a-z]+) ([a-z]+)/\2 \1/'


echo 'abcd_qwer_zxc_123'| sed -r 's/([a-z]+)_([a-z]+)/\2**_**\1/'
echo 'abcd_qwer_zxc_123'| sed -r 's/([a-z]+)**_**([a-z]+)/\2 \1/'

Replacing text between two words

sed -ure 's/**word1**.+?**word2**/**word1**\ replacement\ **word2**/g' -i file

Using variables in sed expressions

When writing bash scripts, we use variables. Sometimes it becomes necessary to use these variables in sed expressions.

The easiest approach is to use double quotes for the wrapper:

sed 's/$var1/$var2/g'

Double quotes won’t work, there are special characters in $var1 or $var2. Your best bet is to exclude variables from the wrapper:

sed 's/''$var1''/''$var2''/g'