Sed

Sed
	Chapter 2. Working in the Console

sed is a stream editor. This sounds scary in the WYSIWYG GUI world to have an editor that does not have a menu or mouse support. see https://devmanual.gentoo.org/tools-reference/sed/index.html or http://www.grymoire.com/Unix/Sed.html.

sed takes all of its inputs from stdin a pipe can be used to feed it. So if called it freezes in the console. Therefore call it best with a file so an EOF comes.
sed can be called from scripts

With sed files can be edited automatically line by line using a single line to call sed. For somebody using sed once a while its syntax and regular expressions is an obstacle, but ask ChatGPT what you want to do as:

Remove commenting lines starting with the # character and the result will be:

sed -i '/^#/d' <filename.txt>

or remove commenting lines with # but do not remove them with ##

sed -i '/^#[^#]/d' <filename.txt>

A good introduction is http://www.grymoire.com/Unix/Sed.html

sed is called as:

sed <options> '<command>' <inputfile>

The single quotes are not necessary but strongly recommended so that the shell never tries to expand it.

sed works as follows:

It takes the input file and the commands from stdin
It reads a line and stores it into its internal pattern buffer
It executes the command to the internal pattern buffer
It puts the result inside the internal pattern buffer to stdout (Note: The input file will not be modified except the -i option is added).
It goes to step 2 until the end of the file has been reached

Filtering certain lines

Of course there are options to limit the commands just to certain lines.

A single line:

sed <options> '<linenumber><command>' <inputfile>

A range of lines:

sed <options> '<first line>,<last line><command>' < inputfile>

A regular expression:

sed '<regular expression><command>' <inputfile>

All lines from a first regular expression to a second regular expression.

sed '<first regular expression>,<second regular expression><command>' < inputfile>

Sed Commands

d delete

p print

s/<find>/<replace>/g find a string and replace it globally

sed 's/\t/ /g' <inputfile> > <output file> replaces tab characters with a blank

sed -i 's/<xxx>/<yyyy>/g' <file> replaces xxx with yyyy in a file

sed -i 's/<xxx>//g' <file> deletes xxx in a file

Sed Options

-n usually all lines if not deleted will be printed. -n does just print the lines that match the -p command.