3 ms·
Try `body`: https://github.com/jeroenjanssens/data-science-at-the-command-line/blob/master/tools/body https://github.com/jeroenjanssens/data-science-at-the-comm
by jeroenjanssens 12y ago
Try `body`: https://github.com/jeroenjanssens/data-science-at-the-command-line/blob/master/tools/body https://github.com/jeroenjanssens/data-science-at-the-comman...
$ echo 'header\ne\nd\na\nb\nc\nb' | body sort | body uniq
header
a
b
c
d
e
- vram22 12y agoHeh, at first I thought you meant a body command like this one, which I've written in the past: $ cat body_test.txt 1 This 2 is 3 a 4 file 5 to 6 test 7 the 8 body 9 command 10 which 11 is 12 a 13 complement 14 to 15 head 16 and 17 tail. $ cat `which body` sed -n $1,$2p $3 $ body 5 10 body_test.txt 5 to 6 test 7 the 8 body 9 command 10 which
- faulteh 12y agoImmediately ^'d this post for it's usefulness, but I put this in my .bashrc instead: function body_alias() { sed -n $1,$2p $3 } alias body=body_alias I used to have little scripts like body in my own ~/bin or /usr/local/bin but I've been slowly moving those to my .bashrc which I can copy to new systems I log on to.
- vram22 12y agoGlad you liked it. Your alias technique is good too. Plus it may save a small amount of time since the body script does not have to be loaded from its file (as in my case) - unless your *nix version caches it in memory after the first time.