A collection of handy Bash One-Liners and terminal tricks for data processing and Linux system maintenance.
Go to file
2016-09-09 18:14:09 +08:00
README.md add sed 2016-09-09 18:14:09 +08:00

Bash-Oneliner

Hi bash learners and bioinformaticans, welcome to Bash Oneliner learning station. I started studying bioinformatics data three years ago, and I was amazed by those single-word bash commands which are much faster than my dull scripts, so i started bash. Not all the code here is oneliner (if the ';' counts..), but i put effort on making them brief and fast.

This blog will focus on bash commands for parsing biological data for dummies like me, most of which are tsv files (tab-separated values); some of them are for Ubuntu system maintaining (for dummies). I have been recording the bash commands on my notebook, but putting them on web will help others and myself to query. I apologize that there won't be any citation for the codes, coz i haven't make any record of it, but they are probably from dear Google and Stackoverflow.

English and bash are not my first language, so... correct me anytime, thank you

In case you would like to check and vote up my questions on Stackoverflow, here's my page: http://stackoverflow.com/users/4290753/once

Here's a more stylish version~ http://onceupon.github.io/Bash-Oneliner/

##Handy Bash oneliner commands for tsv file editing

##Grep #####extract text bewteen words (e.g. w1,w2)

grep -o -P '(?<=w1).*(?=w2)'

#####grep lines without word (e.g. bbo)

grep -v bbo filename

#####grep only one/first match (e.g. bbo)

grep -m 1 bbo filename

#####grep and count (e.g. bbo)

grep -c bbo filename

#####insensitive grep (e.g. bbo/BBO/Bbo)

grep -i "bbo" filename 

#####count occurrence (e.g. three times a line count three times)

grep -o bbo filename 

#####COLOR the match (e.g. bbo)!

grep --color bbo filename 

#####grep search all files in a directory(e.g. bbo)

grep -R bbo /path/to/directory 

or

grep -r bbo /path/to/directory 

#####search all files in directory, only output file names with matches(e.g. bbo)

grep -Rh bbo /path/to/directory 

or

grep -rh bbo /path/to/directory 

#####grep OR (e.g. A or B or C or D)

grep 'A\|B\|C\|D' 

#####grep AND (e.g. A and B)

grep 'A.*B' 

#####grep all content of a fileA from fileB

grep -f fileA fileB 

#####grep a tab

grep $'\t' 

#####grep variable from variable

$echo "$long_str"|grep -q "$short_str"
if [ $? -eq 0 ]; then echo 'found'; fi

//grep -q will output 0 if match found //remember to add space between []!

#####grep strings between a bracket()

grep -oP '\(\K[^\)]+'

#####grep number of characters with known strings in between(e.g. AAEL000001-RA)

grep -o -w "\w\{10\}\-R\w\{1\}"

// \w word character [0-9a-zA-Z_] \W not word character

#####a lot examples here http://www.cyberciti.biz/faq/grep-regular-expressions/

##Sed [back to top]

#####remove lines with word (e.g. bbo)

sed "/bbo/d" filename

#####edit infile (edit and save)

sed -i "/bbo/d" filename

#####when using variable (e.g. $i), use double quotes " " e.g. add >$i to the first line (to make a FASTA file)

sed "1i >$i"  

//notice the double quotes! in other examples, you can use a single quote, but here, no way! //'1i' means insert to first line

#####delete empty lines

sed '/^\s*$/d' 

or

sed 's/^$/d' 

#####delete last line

sed '$d' 

#####delete last character from end of file

sed -i '$ s/.$//' filename

#####add string to end of file (e.g. "]")

sed '$s/$/]/' filename

#####add string to beginning of every line (e.g. bbo)

sed -e 's/^/bbo/' file

#####add string to end of each line (e.g. "}")

sed -e 's/$/\}\]/' filename

#####add \n every nth character (e.g. every 4th character)

sed 's/.\{4\}/&\n/g' 

#####concatenate/combine/join files with a seperator and next line (e.g seperate by ",")

sed -s '$a,' *.json > all.json

#####substitution (e.g. replace A by B)

sed 's/A/B/g' filename 

#####select lines start with string (e.g. bbo)

sed -n '/^@S/p' 

#####delete lines with string (e.g. bbo)

sed '/bbo/d' filename 

#####print every nth lines

sed -n '0~3p' filename

//catch 0: start; 3: step

#####print every odd # lines

sed -n '1~2p' 

#####print every third line including the first line

sed -n '1p;0~3p' 

#####remove leading whitespace and tabs

sed -e 's/^[ \t]*//'

//notice a whitespace before '\t'!!

#####remove only leading whitespace

sed 's/ *//'

//notice a whitespace before '*'!!

#####remove ending commas

sed 's/,$//g' 

#####add a column to the end

sed "s/$/\t$i/"

//$i is the valuable you want to add e.g. add the filename to every last column of the file

for i in $(ls);do sed -i "s/$/\t$i/" $i;done

#####add extension of filename to last column

for i in T000086_1.02.n T000086_1.02.p;do sed "s/$/\t${i/*./}/" $i;done >T000086_1.02.np

#####remove newline\ nextline

sed ':a;N;$!ba;s/\n//g'

#####print a number of lines (e.g. line 10th to line 33 rd)

sed -n '10,33p' <filename

#####change delimiter

sed 's=/=\\/=g'

#Awk [back to top]

#####set tab as field separator

awk -F $'\t'  

#####output as tab separated (also as field separator)

awk -v OFS='\t' 

#####pass variable

a=bbo;b=obb;
awk -v a="$a" -v b="$b" "$1==a && $10=b' filename 

#####print number of characters on each line

awk '{print length ($0);}' filename 

#####find number of columns

awk '{print NF}' 

#####reverse column order

awk '{print $2, $1}' 

#####check if there is a comma in a column (e.g. column $1)

awk '$1~/,/ {print}'  

#####split and do for loop

awk '{split($2, a,",");for (i in a) print $1"\t"a[i]} filename 

#####print all lines before nth occurence of a string (e.g stop print lines when bbo appears 7 times)

awk -v N=7 '{print}/bbo/&& --N<=0 {exit}'

#####print filename and last line of all files in directory

ls|xargs -n1 -I file awk '{s=$0};END{print FILENAME,s}' file'

#####add string to the beginning of a column (e.g add "chr" to column $3)

awk 'BEGIN{OFS="\t"}$3="chr"$3' 

#####remove lines with string (e.g. bbo)

awk '!/bbo/' file 

#####column subtraction

cat file| awk -F '\t' 'BEGIN {SUM=0}{SUM+=$3-$2}END{print SUM}'

#####usage and meaning of NR and FNR e.g. fileA: a b c fileB: d e

awk 'print FILENAME, NR,FNR,$0}' fileA fileB 

fileA 1 1 a fileA 2 2 b fileA 3 3 c fileB 4 1 d fileB 5 2 e

#####and gate

e.g. fileA: 1 0

2 1

3 1

4 0

fileB:

1 0

2 1

3 0

4 1

awk -v OFS='\t' 'NR=FNR{a[$1]=$2;next} NF {print $1,((a[$1]=$2)? $2:"0")}' fileA fileB 

1 0

2 1

3 0

4 0

#####round all numbers of file (e.g. 2 significant figure)

awk '{while (match($0, /[0-9]+\[0-9]+/)){
    \printf "%s%.2f", substr($0,0,RSTART-1),substr($0,RSTART,RLENGTH)
    \$0=substr($0, RSTART+RLENGTH)
    \}
    \print
    \}'

#####give number/index to every row

awk '{printf("%s\t%s\n",NR,$0)}'

#####break combine column data into rows

e.g. seperate

David cat,dog

into

David cat

David dog

detail here: http://stackoverflow.com/questions/33408762/bash-turning-single-comma-separated-column-into-multi-line-string

awk '{split($2,a,",");for(i in a)print $1"\t"a[i]}' file

#####sum up a file (each line in file contains only one number)

awk '{s+=$1} END {print s}' filename

#####average a file (each line in file contains only one number)

awk '{s+=$1}END{print s/NR}'

#####print field start with string (e.g Linux)

awk '$1 ~ /^Linux/'

#####sort a row (e.g. 1 40 35 12 23 --> 1 12 23 35 40)

awk ' {split( $0, a, "\t" ); asort( a ); for( i = 1; i <= length(a); i++ ) printf( "%s\t", a[i] ); printf( "\n" ); }'

##Xargs [back to top]

#####set tab as delimiter (default:space)

xargs -d\t

#####display 3 items per line

echo 1 2 3 4 5 6| xargs -n 3

//1 2 3 4 5 6

#####prompt before execution

echo a b c |xargs -p -n 3

#####print command along with output

xargs -t abcd

///bin/echo abcd //abcd

#####with find and rm

find . -name "*.html"|xargs rm -rf

delete fiels with whitespace in filename (e.g. "hello 2001")

find . -name "*.c" -print0|xargs -0 rm -rf

#####show limits

xargs --show-limits

#####move files to folder

find . -name "*.bak" -print 0|xargs -0 -I {} mv {} ~/old

or

find . -name "*.bak" -print 0|xargs -0 -I file mv file ~/old

#####move first 100th files to a directory (e.g. d1)

ls |head -100|xargs -I {} mv {} d1

#####parallel

time echo {1..5} |xargs -n 1 -P 5 sleep

a lot faster than

time echo {1..5} |xargs -n1 sleep

#####copy all files from A to B

find /dir/to/A -type f -name "*.py" -print 0| xargs -0 -r -I file cp -v -p file --target-directory=/path/to/B

//v: verbose| //p: keep detail (e.g. owner)

#####with sed

ls |xargs -n1 -I file sed -i '/^Pos/d' filename

#####add the file name to the first line of file

ls |sed 's/.txt//g'|xargs -n1 -I file sed -i -e '1 i\>file\' file.txt

#####count all files

ls |xargs -n1 wc -l

#####to filter txt to a single line

ls -l| xargs

#####count files within directories

echo mso{1..8}|xargs -n1 bash -c 'echo -n "$1:"; ls -la "$1"| grep -w 74 |wc -l' --

// "--" signals the end of options and display further option processing

#####download dependencies files and install (e.g. requirements.txt)

cat requirements.txt| xargs -n1 sudo pip install

#####count lines in all file, also count total lines

ls|xargs wc -l

##Find [back to top] #####list all sub directory/file in the current directory

find .

#####list all files under the current directory

find . -type f

#####list all directories under the current directory

find . -type d

#####edit all files under current directory (e.g. replace 'www' with 'ww')

find . name '*.php' -exec sed -i 's/www/w/g' {} \;

if no subdirectory

replace "www" "w" -- *

//a space before *

#####find and output only filename (e.g. "mso")

find mso*/ -name M* -printf "%f\n"

#####find and delete file with size less than (e.g. 74 byte)

find . -name "*.mso" -size -74c -delete

//M for MB, etc

##Loops [back to top] #####while loop, column subtraction of a file (e.g. a 3 columns file)

while read a b c; do echo $(($c-$b));done < <(head filename)

//there is a space between the two '<'s

#####while loop, sum up column subtraction

i=0; while read a b c; do ((i+=$c-$b)); echo $i; done < <(head filename)

#####if loop

if (($j==$u+2))

//(( )) use for arithmetic operation

if [[$age >21]]

// use for comparison

#####for loop

for i in $(ls); do echo file $i;done

##Download [back to top] #####download all from a page

wget -r -l1 -H -t1 -nd -N -np -A mp3 -e robots=off http://example.com

//-r: recursive and download all links on page

//-l1: only one level link

//-H: span host, visit other hosts

//-t1: numbers of retries

//-nd: don't make new directories, download to here

//-N: turn on timestamp

//-nd: no parent

//-A: type (seperate by ,)

//-e robots=off: ignore the robots.txt file which stop wget from crashing the site, sorry example.com

##Random [back to top] #####random pick 100 lines from a file

shuf -n 100 filename

#####random order (lucky draw)

for i in a b c d e; do echo $i; done| shuf

#####echo series of random numbers between a range (e.g. generate 15 random numbers from 0-10)

shuf -i 0-10 -n 15

#####echo a random number

echo $RANDOM

#####random from 0-9

echo $((RANDOM % 10))

#####random from 1-10

echo $(((RANDOM %10)+1))

##Others [back to top] #####remove newline / nextline

tr --delete '\n' <input.txt >output.txt

#####replace newline

tr '\n' ' ' <filename

#####compare files (e.g. fileA, fileB)

diff fileA fileB

//a: added; d:delete; c:changed

or

sdiff fileA fileB

//side-to-side merge of file differences

#####number a file (e.g. fileA)

nl fileA

or

nl -nrz fileA

//add leading zeros

#####combine/ paste two files (e.g. fileA, fileB)

paste fileA fileB

//default tab seperated

#####reverse string

echo 12345| rev

#####read .gz file without extracting

zmore filename

or

zless filename

#####run in background, output error file

some_commands  &>log &

or

some_commands 2>log &

or

some_commands 2>&1| tee logfile

or

some_commands 2>&1 >>outfile

//0: standard input; 1: standard output; 2: standard error

#####send mail

echo 'heres the content'| mail -A 'file.txt' -s 'mail.subject' me@gmail.com

//use -a flag to set send from (-a "From: some@mail.tld")

#####.xls to csv

xls2csv filename

#####append to file (e.g. hihi)

echo 'hihi' >>filename

#####make BEEP sound

speaker-test -t sine -f 1000 -l1

#####set beep duration

(speaker-test -t sine -f 1000) & pid=$!;sleep 0.1s;kill -9 $pid

#####history edit/ delete

~/.bash_history

or

history -d [line_number]

#####get last history/record filename

head !$

#####clean screen

clear

or

Ctrl+l

#####send data to last edited file

cat /directory/to/file
echo 100>!$

#####run history number (e.g. 53)

!53

#####run last command

!!

#####run last command that began with (e.g. cat filename)

!cat

or

!c

//run cat filename again

#####extract .xf

1.unxz filename.tar.xz
2.tar -xf filename.tar

#####install python package

pip install packagename

#####Download file if necessary

data=file.txt
url=http://www.example.com/$data
if [! -s $data];then
    echo "downloading test data..."
    wget $url
fi

#####wget to a filename (when a long name)

wget -O filename "http://example.com"

#####wget files to a folder

wget -P /path/to/directory "http://example.com"

#####delete current bash command

Ctrl+U

or

Ctrl+C

or

Alt+Shift+#

//to make it to history

#####add things to history (e.g. "addmetohistory")

#addmetodistory

//just add a "#" before~~

#####sleep awhile or wait for a moment or schedule a job

sleep 5;echo hi

#####count the time for executing a command

time echo hi

#####backup with rsync

rsync -av filename filename.bak
rsync -av directory directory.bak
rsync -av --ignore_existing directory/ directory.bak
rsync -av --update directory directory.bak

//skip files that are newer on receiver (i prefer this one!)

#####make all directories at one time!

mkdir -p project/{lib/ext,bin,src,doc/{html,info,pdf},demo/stat}

//-p: make parent directory //this will create project/doc/html/; project/doc/info; project/lib/ext ,etc

#####run command only if another command returns zero exit status (well done)

cd tmp/ && tar xvf ~/a.tar

#####run command only if another command returns non-zero exit status (not finish)

cd tmp/a/b/c ||mkdir -p tmp/a/b/c

#####extract to a path

tar xvf -C /path/to/directory filename.gz

#####use backslash "" to break long command

cd tmp/a/b/c \
> || \
>mkdir -p tmp/a/b/c

#####get pwd

VAR=$PWD; cd ~; tar xvf -C $VAR file.tar

//PWD need to be capital letter

#####list file type of file (e.g. /tmp/)

file /tmp/

//tmp/: directory

#####bash script

#!/bin/bash
file=${1#*.}

//remove string before a "."

file=${1%.*}

//remove string after a "."

#####search from history

Ctrl+r

#####python simple HTTP Server

python -m SimpleHTTPServer

#####variables

{i/a/,}

e.g. replace all

{i//a/,}

//for variable i, replace all 'a' with a comma

#####read user input

read input
echo $input

#####generate sequence 1-10

seq 10

#####sum up input list (e.g. seq 10)

seq 10|paste -sd+|bc

#####find average of input list/file

i=`wc -l filename|cut -d ' ' -f1`; cat filename| echo "scale=2;(`paste -sd+`)/"$i|bc

#####generate all combination (e.g. 1,2)

echo {1,2}{1,2}

//1 1, 1 2, 2 1, 2 2

#####generate all combination (e.g. A,T,C,G)

set = {A,T,C,G}
group= 5
for ((i=0; i<$group; i++));do
    repetition=$set$repetition;done
    bash -c "echo "$repetition""

#####read file content to variable

foo=$(<test1)

#####echo size of variable

echo ${#foo}

#####echo tab

echo -e ' \t '

#####array

declare -A array=()

#####send a directory

scp -r directoryname user@ip:/path/to/send

#####split file into lines (e.g. 1000 lines/smallfile)

$ split -d -l 1000 bigfilename

##System [back to top]

#####snapshot of the current processes

ps 

#####check graphics card

lspci

#####show IP address

$ip add show

or

ifconfig

#####check system version

cat /etc/*-release

#####Linux Programmer's Manuel: hier- description of the filesystem hierarchy

man hier

#####list job

jobs -l

#####export PATH

export PATH=$PATH:~/path/you/want

#####make file execuable

chmod +x filename

//you can now ./filename to execute it

#####list screen

screen -d -r

#####echo screen name

screen -ls

#####check system (x86-64)

uname -i

#####surf the net

links www.google.com

#####add user, set passwd

useradd username
passwd username

#####edit variable for bash, (e.g. displaying the whole path)

1. joe ~/.bash_profile 
2. export PS1='\u@\h:\w\$' 

//$PS1 is a variable that defines the makeup and style of the command prompt

3. source ~/.bash_profile

#####edit environment setting (e.g. alias)

1. joe ~/.bash_profile
2. alias pd="pwd" //no more need to type that 'w'!
3. source ~/.bash_profile

#####list environment variables (e.g. PATH)

$echo $PATH

//list of directories separated by a colon

#####list all environment variables for current user

$env

#####show partition format

lsblk

#####soft link program to bin

ln -s /path/to/program /home/usr/bin

//must be the whole path to the program

#####show hexadecimal view of data

hexdump -C filename.class

#####jump to different node

rsh node_name

#####check port (active internet connection)

netstat -tulpn

#####find whick link to a file

readlink filename

#####check where a command link to (e.g. python)

which python

#####list total size of a directory

du -hs .

or

du -sb

#####copy directory with permission setting

cp -rp /path/to/directory

#####store current directory

pushd . $popd ;dirs -l 

#####show disk usage

df -h 

or

du -h 

or

du -sk /var/log/* |sort -rn |head -10

#####show current runlevel

runlevel

#####switch runlevel

init 3 

or

telinit 3 

#####permanently modify runlevel

1. edit /etc/init/rc-sysinit.conf 
2. env DEFAULT_RUNLEVEL=2 

#####become root

su

#####become somebody

su somebody

#####report user quotes on device

requota -auvs

#####get entries in a number of important databases

getent database_name

(e.g. the 'passwd' database)

getent passwd

//list all user account (all local and LDAP) (e.g. fetch list of grop accounts)

getent group

//store in database 'group'

#####little xwindow tools

xclock
xeyes

#####change owner of file

chown user_name filename
chown -R user_name /path/to/directory/

//chown user:group filename

#####list current mount detail

df

#####list current usernames and user-numbers

cat /etc/passwd

#####get all username

getent passwd| awk '{FS="[:]"; print $1}'

#####show all users

compgen -u

#####show all groups

compgen -g

#####show group of user

group username

#####show uid, gid, group of user

id username

#####check if it's root

if [$(id -u) -ne 0];then
    echo "You are not root!"
    exit;
fi

//'id -u' output 0 if it's not root

#####find out CPU information

more /proc/cpuinfo

or

lscpu

#####set quota for user (e.g. disk soft limit: 120586240; hard limit: 125829120)

setquota username 120586240 125829120 0 0 /home

#####show quota for user

quota -v username

#####fork bomb

:(){:|:&};:

//dont try this at home

#####check user login

lastlog

#####edit path for all users

joe /etc/environment

//edit this file

#####show running processes

ps aux

#####find maximum number of processes

cat /proc/sys/kernal/pid_max

#####show and set user limit

ulimit -u

#####which ports are listening for TCP connections from the network

nmap -sT -O localhost

#####print out number of cores/ processors

nproc --all

#####check status of each core

  1. top
  2. press '1'

=-=-=-=-=-A lot more coming!! =-=-=-=-=-=-=-=-=-=waitwait-=-=-=-=-=-=-=-=-=-