Useful SED commands - Linux SED

    While working in Linux, SED always plays important role in text processing. Here we've given some common examples of SED usage. Hope it will be a life saving that comes in handy!!!!!

  Let us know your shortcuts/ commands in comment section below that saved your time...

Linux Commands

• Print the number of lines in myfile.txt

sed -n '$=' myfile.txt

            -n switch tells SED not to print all lines, which is its default action.

• Print lines 5 through 10 from myfile.txt

sed -n '5,10p' myfile.txt

• Print all except lines 20 through end of file

sed '20,$d' myfile.txt

• Append line numbers to the file. (= prints line number & new line)

sed '=' myfile.txt | sed 's/\n/) /'

•  Display lines 5-7 and 10-13 from myfile.txt

sed -n -e '5,7p' -e '10,13p' myfile.txt

• Display all lines containing 'start'

sed -n '/start/ p' myfile.txt

           

• Add a line after match

sed '/start/ a “MATCH FOUND”' myfile.txt

• Add a line before match

sed '/start/ i “MATCH FOUND”' myfile.txt

• Change a matching line

sed '/start/ c “MATCH FOUND”' myfile.txt

• Display all lines containing beginning 'start'

sed -n '/^start/ p' myfile.txt

• To print all the lines after finding the first occurrence of the pattern.

sed -n '/^start/,$ p' myfile.txt

• To print 2 lines after finding the first occurrence of the pattern.

sed -n '/^start/, +2 p' myfile.txt

• Adds some text before line numbers 3 to 5

sed '3,5 s/^/#/' myfile.txt

• To remove blank lines ( ^ denotes start of the line, $ denotes end of line)

sed '/^$/d' myfile.txt

• To remove lines starting with 'start'

sed '/^start/d' myfile.txt

• To remove lines ending with 'start'

sed '/start$/d' myfile.txt

• To remove lines containing 'start' and 'end'    (.* stands for any character

sed '/.*start.*end.*/d' myfile.txt

• To remove lines starting or ending with '#' or empty lines

sed '/^#\|#$\|$^/d' myfile.txt

• To replace multiple blank spaces with a single space

sed '/  */ /g' myfile.txt

• To insert one blank line every line

sed G myfile.txt

• To insert two blank line every line

sed 'G;G' myfile.txt

• To replace every instance of the word version with story in myfile.txt, do:

sed 's/version/story/g' myfile.txt

• To ignore character case, use 'gi' instead of 'g'

sed 's/version/story/gi' myfile.txt

• To replace words only within a line range (30 through 40, for example)

sed '30,40 s/version/story/g' myfile.txt

• To replace words except within a line range (30 through 40, for example)

sed '30,40!s/version/story/g' myfile.txt

• To replace only 2^nd match of the search pattern

sed 's/version/story/2' myfile.txt

• Emulating dos2unix with inline editing, by removing (^M)  from the file created in Windows, and save the output in same file.

sed -i 's/\r//' myfile.txt

• Replacing words only if a separate match is found. For example replace to replace 'start' with 'stop' only if the word services is found in the same line

sed '/services/ s/start/stop/g' myfile.txt

• Performing two or more substitutions at once in the same file

sed -i 's/that/this/gi;s/line/verse/gi' myfile.txt

• To replace a word beginning with uppercase or lowercase with another word, example word 'zip' or 'Zip' with 'rar' in myfile.txt

sed 's/[Zz]ip/rar/g' myfile.txt

• To replace a pattern with some text including search pattern.

sed 's/Zip/Create & and send mail/g' myfile.txt

The above command does Zip to  → Create Zip and send mail to

• Do replacement and write the changes done to a seperate file.

sed 's/Zip/Create & and send mail/w changes.txt' myfile.txt

            changes.txt will have  “Create Zip and send mail to”

• Transform function works similar to replacement, instead it works on the character replacement in place of pattern. For example replace  h→H, s→S, i→I, x→W

     sed 'y/hsix/HSIY/' myfile.txt

     sed '2y/hsix/HSIY/' myfile.txt  => performs in 2^nd line 

• Read function, prints the contents of the specified file in the next line where the pattern is found. For example, when 'NAME:' is found, add the contents of 'namelist.txt'

 sed '/NAME:/r namelist.txt' myfile.txt

• Switching pairs of words. Lets say, to switch  'first, last' with 'last, first' in the lines starting with 'NAME'

sed 's/^NAME: \(\w*\), \(\w*\)/NAME: \2, \1/g' myfile.txt

            or

sed 's/\(\w\+\), \(\w\+\)/\2, \1/g' myfile.txt

            or

sed 's@\(\w\+\), \(\w\+\)@\2, \1@g' myfile.txt

 

Valid delimiters can be  @ | ! ^

with 's!/dev/null!/dev/stdout!gp'  → /dev/null becomes /dev/stdout

useful when the string has '/' characters.

Below table provides the short codes for pattern matching. It is very similar to Perl RegEx. 

Code	Description
\w	Alphanumeric Characters including _
\W	Non-Alphanumeric Characters
\s	White Space
\S	Non-White Space
\d	Digits
\D	Non-Digits
.	Any character
*	Zero or more occurrence
\+	One or more occurrence
\?	Zero or one occurrence
\{m\}	Exactly m number of occurrence
\{m,n\}	Within m to n occurrence
\{m,\}	Greater or equal to  m  occurrence