sed¶

I can't believe it's not butter

in-place editing¶

this will replace the files and create a .bak backup file

sed -i.bak -e 's/foo/bar/g' filename

without backup:

sed -i '' -e 's/foo/bar/g' filename

remove all lines except matching one¶

let's say you've got this file with sha hashes and you want to verify just one file of it (and you're using a version of sha256sum which can't just ignore/skip missing files):

8edf5e31e28c5e56df3f87334d193cc777f241481ba091930ec53220b01c135d  yt-dlp.exe
e35bc18f766fd61ed46a42913f89680d5ea1d714135898eb6df5cd9248213ccc  yt-dlp_x86.exe
25cc412d1d3c0725a1f2f5b7e4682f6fb40e6d15f7024e96f7afd572e9919535  yt-dlp
ae08e0b56fea59a8bfdadacd92eddc9bdfdc1473199178cb4e31bacfd991864a  yt-dlp.tar.gz
5bd59c0b95740cfd667b12c627b93c0031e1d4e2fbb3e09b7ede293b48757bee  yt-dlp.zip

I just want the line containing the yt-dlp hash, not the .zip, .tar.gz etc.

sed -i '/yt-dlp$/!d' filename.txt

remove after match¶

example: foo:1234, I want to get rid of :1234

echo foo:1234 | sed 's/\:.*//'

remove empty lines¶

sed '/^$/d'

append to end of line¶

I want to add a comma , to the end of every line

sed 's/$/,/'

remove spaces¶

sed 's/ //'

replace multiple patterns in one go¶

sed -e 's/foo/bar/g' -e 's/bat/baz/g' filename