awesome-terminal-hacks

A repository consisting of useful commands required in daily tasks to reduce stackoverflow searches.

Just click on the right of the codeblock to copy the command and use it directly in your work.

Task

Find the list of all extensions in a folder and subdirectories.

find . -type f | perl -ne 'print $1 if m/\.([^.\/]+)$/' | sort -u

Find the list of all extensions in a folder and its subdirectories with counts

find . -type f | sed -e 's/.*\.//' | sed -e 's/.*\///' | sort | uniq -c | sort -rn

Alternate command for symlinks too

find . -type d -exec echo dirs \; -o -type l -exec echo symlinks \; -o -type f -links +1 -exec echo hardlinks \; -o -type f -exec echo files \; | sort | uniq -c

Find files greater than a particular size in linux

 find . -type f -size +4G

find files less than a particular size in Linux

find . -type f -size -4G

NVIDIA command - check the GPU resource usages

nvidia-smi && nvidia-smi pmon -c 1 | awk -F ' ' '{print $2}' | xargs -0 > t1 && export i=0; export l2="-"; while IFS='' read -r line;do if [[ "$i" -lt 2 || -z "$line" || "$line" == $l2 ]];then echo $i; else ps -p "$line" -o pid,vsz=MEMORY -o user,group=GROUP -o comm,args=ARGS | awk '{for ( x=1 ; x<=1 ; x++ ) { printf("%s\t",$x) } for (x=2;x<=2;x++) { if(NR>1) { printf("%13.2fMb\t",hr=$x/1024) } else { printf("\t%s\t",$x) } } for ( x=3 ; x<=NF ; x++ ) { printf("%s ",$x) } print "" }'; fi; i=$((i+1)) ;done < t1 && rm t1

NVIDIA command - Fix `Failed to initialize NVML: Driver/library version mismatch`

Please note, this command requires sudo and will kill all processes using the GPU.

# Kill all processes using the GPU
sudo lsof | grep /dev/nvidia | awk '{print $2}' | xargs -r kill
# Give processes a moment to terminate
sleep 1
# Forcibly kill processes
sudo lsof | grep /dev/nvidia | awk '{print $2}' | xargs -r kill -9
# Unload old kernel modules (drivers)
sudo rmmod nvidia_uvm nvidia_drm nvidia_modeset nvidia
# New kernel modules will automatically reload
nvidia-smi

Perform operation on all files in a path -- example here is to remove _ from filenames

for FILENAME in *; do FILENAME1=${FILENAME#*_}; mv $FILENAME $FILENAME1; done

Note - this above for loop can be used to do many stuff accordingly. Useful to have this in mind

Swap columns in a file - can be modified as per your needs

awk 'BEGIN{FS=OFS=","}{$0=$2 FS $2 FS $1}1' <your comma delimited file>

Note - read awk documentation to perform lots of crazy commmands based on the above template. Learn awk - its very useful nifty tool and no installation needed on any Linux OS.

Copy fastest using tar

tar cf - * | mbuffer -m 1024M | ssh user@host '(cd /home/path; tar xf -)'

Copy fast using rsync with OSX

rsync -rltv --progress --human-readable --delete --iconv=utf-8-mac,utf-8 -e 'ssh -T -c [email protected] -o Compression=no -x' <local_mac_source> <remote_linux_dest>

rsync -rtvvv --numeric-ids --delete --progress -e 'ssh -T -c [email protected] -o Compression=no -x'  <serve> <local>

https://gist.github.com/KartikTalwar/4393116

Copy k files from one folder to another (assumes you are running command from current folder)

find . -maxdepth 1 -type f | head -k | xargs cp -t <destination folder>

Extract wav audio files form an mp4 extension of video - crude way - more customization is possible.

for f in *.mp4;do ffmpeg -i "$f" "${f%mp4}wav";done

Replace spaces with underscores -- very irritating to handle spaces in filenames -- easy tool to fix that so your code doesnt need to handle.

for file in */*; do echo $file; mv "$file" $(echo $file | sed 's/ /_/g'); done

Quickly get sum of a file consisting of numbers

awk 'BEGIN {FS=" "};{sum+=$NF} END {print sum}' <your file>

Common lines between 2 files in Linux

comm -12 <(sort file1) <(sort file2)

grep -Fxf file1 file2

both grep and comm has same execution time roughly for relatively small txt files less than 50mb. Post that grep is better for larger files.

Get Lines unique to file1 in linux

comm -23 <(sort file1) <(sort file2)

Get lines unique to file2 in linux

comm -13 <(sort file1) <(sort file2)

Ignore case in comm command by passing comm -i

Get unique lines in your file (sort if not sorted)

sort file1 | uniq

Get number of times unique lines appear in a file (sort if not sorted)

sort file1 | uniq -c

Only print the repeated lines once in a file. (sort if not sorted) (basically count > 1 lines)

sort file1 | uniq -d

Only print the lines which are not repeated in a file. (sort if not sorted) (basically count == 1 lines)

sort file1 | uniq -u

Run uniq command on a given field (below command will run uniq only field 1 - useful in delimited files)

uniq -f 1 file1

Note: above command is sample and can be used with previous commands in combination to get what you want from a particular field.

View range of lines from a file

sed -n '1,5p' file1

View full file except range from a file

sed '3,10d' file1

View non consecutive lines and ranges from a file - example below shows line 2-4 and 8-12

sed -n -e '2,4p' -e '8,12p' file1

Replace word1 by word2 in a file (use -i to overwrite file)

sed 's/word1/word2/g' file1

To ignore case of replacement

sed 's/word1/word2/gi' file1

Rename all files in directory - Removes .wav for example in below

for i in `ls`; do echo $i; new_name=`echo $i | sed s/.wav//`; echo $new_name; mv $i $new_name; done

Replace word1 by word2 in a range in a file - example for 5-10 lines replacement

sed '5,10 s/word1/word2/g' file1

dos2unix (if not available) - convert file format for unix with sed

sed -i 's/\r//' file1

In place replacement for word1 by word 2 with sed and backing up original

sed -i'.orig' 's/word1/word2/gi' file1

Replace word1 by word2 only if word3 present in line - with sed

sed '/word3/ s/word1/word2/g' file1

Multiple substitutions in sed - inplace example

sed -i 's/word1/word2/gi;s/word3/word4/gi' file1

Print specific fields from a commands output - who can be any other command

who | awk '{print $1, $2}'

Omit fields when printing from a file into terminal

awk -F":" '{$2="";$3="";print}' file1

Remove empty lines from a file in linux

awk '/^[ \t]*$/{next}{print}' file1

Get number of lines different between two files

diff -y --suppress-common-lines file1 file2 | wc -l

Get number of fields on each line in a file

awk '{print NR,"-->",NF}' file1

Print all lines greater than a given length - in a file/echoed output

echo "Some long string...." | awk 'length($0) > 5'

Get word frequency in a document using awk

awk 'BEGIN {FS="[^a-zA-Z]+" } { for (i=1; i<=NF; i++) words[tolower($i)]++ } END { for (i in words) print i, words[i] }' file1

Rename all files in a path with awk - example .WAV to .wav

ls *.WAV | awk '{ printf("mv \"%s\" \"%s\"\n", $0, tolower($0)) }'

Add line number before lines in a file

awk '{ print FNR ". " $0 ;next}{print}' file1

Replace the Nth occurance of a string in vim -- very useful stuff

:%s/\(\zsPATTERN.\{-}\)\{N}/REPLACE/

Copy based on awk in vim -- you can basically copy with delimiters - useful

:%!awk '{print $0" "$1}'

Get full path of all files in a directory and dump to a file.

find "$PWD" > filename

Copy a file k times

for i in {1..k}; do cp file "file$i"; done

Split a folder into k parts

for i in `seq 1 k`; do mkdir -p "folder$i"; find . -type f -maxdepth 1 | head -n num_files_you_WANT | xargs -i mv "{}" "folder$i"; done

Speed up listing of files in a path consisting of many many files. (difference noticable after 50k files with standard ls command)

ls -1f <your path>

For adventerous folks - compile this c code - https://github.com/ChristopherSchultz/fast-file-count Its faster than ls -1f by about 6x on a folder containing 1 million files. Note - speedup depends on many factors such as your kernel version, operating version, gcc version -- but definitely the fast-file-count c code is faster. I have it in my home path always.

du - is very slow - to get the filesize of a folder with many files - What to do??

Solution - diskus

Fast Commands with gnu parallel installed (https://anaconda.org/conda-forge/parallel)

You will save a lot of time with this!!! Use parallel with rm with caution. I messed up a linux system due to this error on my personal laptop once.

Convert jpeg to png in a path with parallel

Slow command

for i in *jpeg; do convert $i $i.png ; done

Fast command with parallel

find . -name "*jpeg" | parallel -I% --max-args 1 convert % %.png

Run a bunch of commands - Keep all commands disjoint (any command should not depend on output of previous command)

Say you have k cores on a machine, create a file called - jobs.txt or something Sample jobs.txt

command1
command2
command3
command4
and so on

Now with k cores run the jobs.txt file (try to run with k-1 commands in parallel)

parallel -- jobs k-1 < jobs.txt

Parallel copy files in linux from a given path containing many files say with 10 cores

parallel -j10 cp {} output_path/ ::: *

Parallel calling with a bash script

Your bash script should have a command per line. This saves a LOT of time!!!

parallel -a your-script-here.sh

Speed up file transfer from pathA (source computer) to pathB (remote computer)

Generate a transfer.log file in /tmp for all the folders/files from a source path you want to get copied

cat /tmp/transfer.log | parallel --will-cite -j 5 rsync -avzm --progress --relative --stats --safe-links --ignore-existing --human-readable {} dest_host:/data/

MacOS Homebrew upgrade - use brewski

alias brewski='brew update && brew upgrade && brew cleanup; brew doctor'

Parallel Tutorial

https://www.gnu.org/software/parallel/parallel_tutorial.html

Check commands from the current user that are running on machine

ps -e -o user= | sort | uniq -c | sort -n

List all open jobs from a user and kill all of them

ps -u username -o pid= | awk '{print $1}' | xargs -I {} kill -9 {}

List number of files in directory and all its subdirectories

for subdir in `ls`; do echo $subdir; ls $subdir/ | wc -l ; done

Screen scroll back

echo 'termcapinfo xterm* ti@:te@' >> ~/.screenrc

Note

If you know a better way to do the tasks Or other tasks which can be added here - kindly create a pull request for the repository.

Many of the commands here are created as per my needs and can be easily modified to ones suited task.

This repository will be continuously updated over time.

Star the repository/share it if this helps reduce your time to make searches/helps in your work.

ankitshah009 / awesome-terminal-hacks Goto Github PK