sed and awk are powerful tools for text stream processing. This article provides a quick-start reference for the sed command:

Delete Command

Delete the Nth line

sed 3d abc.txt

This command deletes the third line.

Delete every other line

	sed '1~2d' abc.txt
	sed '2~2d' abc.txt

The first command deletes odd-numbered lines, and the second deletes even-numbered lines.

Delete lines from N to M

	sed '5,11d' abc.txt

Deletes lines 5 through 11.

Delete the last line

	sed '$d' abc.txt

Delete lines matching a pattern

	sed /unix/d abc.txt

This command deletes all lines containing the unix keyword.

Delete a matching line and the N lines after it

	sed '/unix/,+3d' abc.txt

This command deletes the line containing the unix keyword plus the next 3 lines.

Delete from a matching line to the end of the file

	sed '/unix/,$d' abc.txt

This command deletes the line containing the unix keyword and all lines after it to the end of the file.

Delete all blank lines

	sed '/^$/d' abc.txt

This command deletes all blank lines in abc.txt.

Delete all comment lines

Here we assume lines are commented with the # symbol.

	sed '/#.*/d' abc.txt

This command deletes all comment lines.

Substitution Command

s stands for substitution, and / is the delimiter.

Single substitution

	sed 's/old/new/'

This command replaces the first occurrence of ‘old’ with ’new’.

Global substitution

	sed 's/old/new/g'

This command replaces all occurrences of ‘old’ with ’new’.

Substitute a piped data stream

	echo "old" | sed 's/old/new/'

This command outputs ’new'.

Substitute a file’s data stream

	sed 's/old/new/' filename

This command replaces ‘old’ with ’new’ and prints the result to the console.

Substitute a file’s data stream and save it

	sed -i 's/old/new/' filename

The -i option replaces ‘old’ and saves the result back to the file.

• Tags:
• Shell
• Linux
• Text Processing