Table of Contents

• Introduction
• Several commands that I used

Introduction

sed is a Unix stream editor¹ that performs basic text transformations on an input stream (a file or input from a pipeline). It can perform operations such as search and replace, insertion and deletion, and more, on individual lines of a file or a range of lines specified in the input. The results are then written to the output stream, which can be redirected to a file, to another command in a pipeline, or to the terminal.

The general syntax for sed command:

sed [options] 'command' input-file

• options: Optional flags that modify the behavior of the sed command. Some common options include -n for suppressing default output, -e for executing multiple commands, and -i for performing in-place editing of files.
• command: A single or a set of commands that specify the transformations to be applied to the input stream. Commands must be enclosed in single quotes ('') to prevent shell expansion.
• input-file: The input file or files to be processed by sed. If no file is specified, sed will read from the standard input stream.

For example, the following command will replace all occurrences of the word “apple” with “orange” in the file fruits.txt:

sed 's/apple/orange/g' fruits.txt

Several commands that I used

• To edit file in place use option -i with sed.

• Backup and edit input file in-place use option -i.bkp with sed.

• Replace a string in multiple files

  sed -i 's/foo/bar/g' *.dat
  

  This will replace word “foo” to “bar” in all *.dat files present in PWD.

  sed -i 'Ns/foo/bar/g' *.dat
  

  Here N is the Nth line of the text file. This command will find “foo” in Nth line of text fine and replace with “bar”. This is not going to check any other line except Nth line.

  To use variables in the sed command use double quotes

  sed -i "s/$var1/ZZ/g" "$file"
  

  Reference: https://askubuntu.com/a/76842

• Replace several character from a string

  echo ramkrishna | sed 's/ram//g'
  

  or if want to replace several character from string, then

  echo "ram> krishna <Shar,ma" | sed 's/[<>,]//g'
  

  The above command will remove > or < or , from the string.

• Show first line number 1-20 in a text file

  sed -n 1,10p FileName.txt
  

• Append a line after text match

  sed '/<search-pattern>/a <Text-To-Append>' FileName.dat
  

• Append a line before text match

  sed '/<search-pattern>/i <Text-To-Append>' FileName.dat
  

  Example:

  $ cat test
  
  foo
  bar
  option
  baz
  

  sed '/option/a insert text here' test
  

  Output

  foo
  bar
  option
  insert text here
  baz
  

  Ref: https://unix.stackexchange.com/questions/121161/how-to-insert-text-after-a-certain-string-in-a-file

• Remove line of text from multiple files

  sed '/line of text/d' *.txt
  

  If you want to edit in place use option -i (edit in place). Or if you want to keep backup of old file then use -i.bak flag. If not simply put it in a bash loop:

  for file in $(ls *.txt)
  do
      sed '/line of text/d' $file > $file.new_file.txt
  done
  

  Ref: https://stackoverflow.com/questions/1182756/remove-line-of-text-from-multiple-files-in-linux

• Copy part of file to another file

  sed -n 16,80p file1 > new_file
  

• Add prefix/suffix to the begining of each line

  sed -i -e 's/^/prefix/' file1     # ^ stands for the begining of line
  sed -i 's/$/:80/' file.txt        # $ stands for the end of line