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/$/]/' filename
add string to end of each line (e.g. "}")
sed -e 's/$/\}\]/' 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
(command here) 2>log &

or

(command here) 2>&1| tee logfile

or

(command here) 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-=-=-=-=-=-=-=-=-=-