Ask HN: What are your favorite one-liner shell commands you use?
There are plenty of lists online with fancy or obscure one-liners, but I’m curious about the ones people actually use day to day.
using pv (pipeviewer) instead of cat to get a progress bar when grepping huge files.
Using httpie instead of curl so I can remember the flags.
The power of find -exec to run commands on a lot of specific files. The nice part is you can run it without exec first to see if you have the right set of files.
If you do a loop you can echo -e “$somevar\r” and then each write will overwrite the previous line so you screen doesn’t fill up but you do get a feel for the progress (to make it nice you need to pad with spaces, google for echo carriage return to learn more)
How does this work?
https://www.geeksforgeeks.org/linux-unix/pv-command-in-linux...
df -h /; echo "----"; for fattable in $(find /var/lib/mysql/ -name *.ibd -size +1G -exec ls -lh {} \; | awk '{ print $9 }' );do echo BEFORE " " $(ls -lh $fattable| awk '{ print $5" " }'); db=$(echo $fattable| cut -d/ -f5); otable=$(echo $fattable| cut -d/ -f6| cut -d. -f1); echo mysql -qbse \"use $db \; optimize table $otable\;\"|bash;echo AFTER" "" " $(ls -lh $fattable| awk '{ print $5" " }');echo "----" ;done; echo "----"; df -h /
apachectl fullstatus | grep ^[0-9]| awk '{ print $12" "$14 }' | sort | uniq -c| sort -n |tail
tail -100000 /var/log/nginx/access.log | sed "/$(hostname -i)/d" | awk '{ print $1" "$7 }' | sort | uniq -c | sort -n | tail -050
find . -mtime -1 | cut -d/ -f2 |uniq
pkill lsphp; sleep 2; while true; do sleep .4; strace -p $(ps aux | grep [i]ndex| awk '{ print $2 }' | head -01); done
for file in /proc/*/status ; do awk '/VmSwap|Name/{printf $2 " " $3}END{ print ""}' $file 2>/dev/null; done| awk '{ print $2" "$3" "$1 }' | sort -n | tail -20
find . -size +200M -exec ls -lh {} \; | awk '{ print $5" "$9 }'
awk '{ print $4 }'
alias ll=“ls -l”
alias lln=“ls -lrt”
Apart from this, I have a few aliases defined to get the size of specific folders and their subfolders (using ‘du -h’ for human readable sizes). The aliases are named like “duh”, “dut” and so on.
<make/build script> 2>&1 | tee build.log | grep <whatever>
2) Do everything within GNU Screen window with CTRL-A + Shift-H to log all output to logfile i.e. "screenlog.<window num>".
Both lifesavers when working with multiple systems and codebases.
3) Always use "set -o vi" with bash so that i can use vi/vim keybindings across everything.
tmux users can use it's `capture-pane` command, either before or after the fact if history is set big enough. There are several helpful flags worth researching.
e.g. `tmux capture-pane -pS - > ~/tmux.log` in a shell to save the history of that pane, or just `prefix+: capture-pane ...` from within tmux
More to the letter of the question, I use "cd -" frequently, "ps -e | grep some_process_i_would_like_the_pid_for", and while I don't use it frequently, I didn't know about "ssh-copy-id" for a long time, and would do it manually with a text editor in the past. Sorry if they are not sufficiently fancy - but for things to get used day to day for me, they will need to be short and sweet.
Install Tailscale on your machines, and your work machine is always just one ssh command away.
git grep foo | cut -d: -f1 | uniq rg -l foo | uniqFor complex one liners, I keep them in a simple txt file with a brief description so I can find them later when I need to use them again.
```alias hstat="curl -Lo /dev/null --silent --head --write-out '%{http_code}\n'" $1 #HTTP status code```
example:
```$ hstat google.com 200```
vimdiff <( jq < /tmp/file1.json ) <( jq < /tmp/file2.json )Great simple test for network speed on a box without speedtest-cli or other tools installed.
docker ps
docker kill
git switch
git commit
git push