Some Useful Unix Commands

For those new to Linux/Unix, here are some useful commands along with some of their useful options. The brackets [ ] mean that an argument is optional. In the examples, bold type refers to commands that the user types. Unix has many more commands, and these commands have more options than those listed here. For more information use the 'man' command or the 'info' command.

man [options] command

Display a manual page (online help) for a command.
Manpages show the syntax and options for a command, and often they give examples and list similar commands. The command 'info' is also useful.

Useful options:

-f	Display a one-line description of the command; synonym 'whatis'
-k	List the man pages related to the command; synonym 'apropos'

Examples:

Display a man page for the 'ls' command

man ls

Display a one-line description of perl

man -f perl

whatis perl

List all man pages related to LaTeX

man -k latex

apropos latex

pwd

Print working directory (print the name of the directory that you are currently in)

cd [directory_name]

Change working directory

Examples:

Change working directory to your home directory (i.e. go to your home directory)

Change to the subdirectory called 'temp'

cd temp

Change to home directory of user 'smith'

cd ~smith

Change to the public_html subdirectory of your home directory

cd ~/public_html

mkdir [options] directory

rmdir [options] directory

Create and remove directories.

ls [options] [file(s)]

List files (or directories)

Useful options:

-l	long listing
-h	print file sizes in reasonable units, not just bytes
-a	list hidden files (beginning with .)
-R	list directories recursively
-d	don't list contents of directories
-S	sort by size (use with -l for example)
-t	sort by time of last file modification
-X	sort by extension
-F	append a symbol to the filename depending on the file type
--color	use color to distinguish between types of files
--color=never	don't use color
--color=auto	use color if output is going to a terminal

Examples:

List files (in current directory) which end in .txt in the long format

ls -l *.txt

List all files and directories which begin with 't'; don't list contents of directories.

ls -dl t*

List all files in current directory, sorted by modification time, and redirect output to the file listing.txt. Don't attempt to use color. Append to the filename a symbol which depends on the file type.

ls -ltF --color=never > listing.txt

cat [options] file(s)

Concatenate files, sending the output to standard output (typically to the computer screen).

Examples:

Print the file file.txt to the computer screen

cat file.txt

Concatenate files file1.txt, file2.txt, and file3.txt, and put the result in file4.txt

cat file1.txt file2.txt file3.txt > file4.txt

more [options] file(s)

less [options] file(s)

Display contents of files.

Useful commands while viewing a file with 'more' or 'less':

<spacebar>	scroll forward one page
^b (control-b)	scroll backward one page
/<text>	search for string <text>
n	find next occurence of string
N	find previous occurence of string (search for string in opposite direction)*
g	go to the beginning of the file*
G	go to the end of the file*
v	edit the current file with 'vi'*
:n	go to the next file*
:p	go to the previous file*
q	quit

*works in 'less', not 'more'

Example:

Display contents of the .txt files in the current directory

more *.txt

mv [options] source destination

Move or rename file; 'destination' is either a file or a directory

mv [options] source1 source2 ... destination

Move file(s); 'destination' must be the name of a directory

Useful options:

-i	prompt user before overwriting destination file if it exists
-f	don't prompt user about overwriting
-u	if the destination file exists, overwrite it only if the source file is newer

Examples:

Rename the file README to readme.txt

mv README readme.txt

Rename the file README to readme.txt, but if readme.txt already exists be prompted whether you want to overwrite it

mv -i README readme.txt

Move all .html files in the current directory into the public_html subdirectory of your home directory, and be prompted before overwriting any existing files

mv -i *.html ~/public_html

Move the file index.html into your public_html directory. If there is already a file there called index.html, don't overwrite it unless the file that you are trying to move is newer.

mv -u index.html ~/public_html

Move the subdirectories 'raw_data' and 'images' to the subdirectory called 'experiment' in your dv01 directory

mv raw_data images ~/dv01/experiment

cp [options] source destination

Copy file; 'destination' is either a file or a directory

cp [options] source1 source2 ... destination

Copy file(s); 'destination' must be the name of a directory

Useful options:

-i	prompt user before overwriting destination file if it exists
-f	don't prompt user about overwriting
-u	if the destination file exists, overwrite it only if the source file is newer
-r or -R	copy directories recursively (copy directories and their files and subdirectories, etc.)
-p	preserve file attributes, such as file timestamps

Examples:

Copy the file README to readme.txt; preserve file attributes

cp -p README readme.txt

Copy the file README to readme.txt, but if readme.txt already exists be prompted whether you want to overwrite it

cp -i README readme.txt

Copy all .html files in the current directory into the public_html subdirectory of your home directory, and be prompted before overwriting any existing files

cp -i *.html ~/public_html

Copy the file index.html into your public_html directory. If there is already a file there called index.html, don't overwrite it unless the file that you are trying to move is newer.

cp -u index.html ~/public_html

Copy the subdirectories 'raw_data' and 'images' to the subdirectory called 'experiment' in your dv01 directory

cp -r raw_data images ~/dv01/experiment

rm [options] file(s)

Remove (delete) files.
It is generally best to avoid this command late at night and on Friday afternoons.

Useful options:

-i	prompt user before removing the file
-f	don't prompt user
-r	recursively remove subdirectories (and their files and subdirectories, etc) - Use With Caution!

Examples:

Remove files with the .exe extension, but prompt user first

rm -i *.exe

Remove subdirectories 'temp' and 'temp2' and all of their contents

rm -r temp temp2

lpr [options] file(s)

Print files. (Sends files to the print server to be printed.)

Useful option:

-P<printer>

print to the specified printer*

*CABI printers available: hp4050 (Image Analysis Lab), dellcolor (Raj's office), hp4m (CABI West).
If the -P option is not specified, files will be printed to the default printer for the computer that you are using. For example, files printed from a Linux computer in the Image Analysis Lab will be sent to the HP LaserJet 4050 printer by default.

Example:

Print the file 'readme.txt' to the printer 'hp4050'

lpr -Php4050 readme.txt

lpq [options]

Shows status of print queues. (Shows files waiting to be printed.)

Useful options:

-P<printer>	show print queue for the specified printer
-a	lists all print queues

who [options]

List the users who are logged in

Useful option:

-H	Print column headings

ps [options]

List processes

Useful options:

-e	List all processes
-f	Use full format.
-U <user>	Select processes for user <user>
a	All processes on a terminal
u	User oriented format
x	Processes without a controlling tty

Note: The full format (-f option) lists the parent process ID (PPID). The PPID of a process is the PID of the process that spawned it, which is sometimes useful to know.

Examples:

List your processes associated with the current shell, in full format

ps -f

List processes for user 'smith'

ps -U smith

List all processes on the computer in full format, except for system processes

ps -ef | grep -v root

ps aux | grep -v root

top [options]

Monitor processes continuously. Updates every 3 seconds by default. Displays useful information such as total memory, memory in use, % of total memory used by each process, % cpu usage of each process, and process ID (PID) of each process. Type 'h' for help or 'q' to quit. To stop ('kill') a process, type 'k' and enter the PID of the process, then enter the signal for the 'kill' command (see below) or press Enter to use the default signal (terminate).

Useful options:

-d <delay_time>	Specify time between updates
-u <user>	List processes for user <user>
-p<PID>	List process with ID <PID>
-n <num_iterations>	Specify number of iterations before quitting

Examples:

Monitor processes associated with user 'smith'

top -u smith

Monitor processes and update every 10 seconds

top -d 10

kill [options] <PID>

Send a signal (such as terminate) to a process. PID = Process ID (from ps or top command).

Useful options:

-l	List possible signals to send (that's the letter 'l' and not the number '1')
-15 or -TERM	Send the 'terminate' signal (the default signal) to the process*
-9 or -KILL	Send the 'kill' signal to the process*
-1 or -HUP	Send 'hang up' signal to the process (that's the number '1' and not the letter 'l')

*Sometimes a program is not terminated by the -TERM (-15) signal. The -KILL (-9) signal is more likely to kill it, but it might not let the program 'clean up' after itself. So if you need to stop a process, first use the default -TERM signal; if that doesn't work, you can try the -KILL signal.

Examples:

List the different possible signals to send.

kill -l

Stop process 19891. (For example, the program associated with process 19891 is hung.)

kill 19891

Stop process 19891 using the 'kill' signal if the previous command didn't work.

kill -9 19891

find [path...] [expression]

Search for file(s), starting with path(s) given. Searches subdirectories recursively. When a file matches the expression, its path is printed (unless -prune is the only action given in the expression).

Useful options/tests/actions:

-name <filename>	Search for the file <filename>
-type <filetype>	Search for file of type <filetype>. Some possible values of <filetype>: d=directory, l=symbolic link, f=regular file
-exec <command>	Execute the command given
-ok <command>	Like exec but prompt user before executing the command
-mtime <n>	Search for file modified <n> days ago. +n = more than n days ago; -n means less than n days ago.
-mmin <n>	Search for file modified <n> minutes ago. +n = more than n minutes ago; -n means less than n minutes ago.
-newer <file>	Search for file newer than <file>
-prune	When a match is found, do not descend into that directory. (Ignore that directory and its subdirectories.)
-not <expr>	True if <expr1> is false
<expr1> -o <expr2>	Or. If <expr1> is false, test <expr2>

Examples:

Search current directory and all subdirectories for all .c files last modified less than 3 days ago

find . -name "*.c" -mtime -3

Search for all files without the .txt extension

find . -not -name "*.txt"

Search your home directory and all subdirectories for a program called 'myprogram.c' which is newer than the one in the directory ~/src/myprogram

find ~ -name myprogram.c -newer ~/src/myprogram/myprogram.c

Search home directory and all subdirectories for the file 'readme.txt'. Skip directories called 'Backup'.

find ~ -name Backup -type d -prune -o -name readme.txt

grep [options] <pattern> [files]

Search inside files for the given pattern. Outputs each line that contains a match. See the manpage for grep for more information about patterns.

Useful options:

-i	Ignore case
-r or -R	Search files in directories recursively
-v	Invert match: select nonmatching lines

Examples:

Search all .c files in the current directory for the string 'flibbertigibbet', ignoring case

grep -i flibbertigibbet *.c

Search all subdirectories recursively for .txt files containing the string 'It was a dark and stormy night'

grep -r "It was a dark and stormy night" *.txt

du [options] [files]

Summarize disk usage for each file and directory. Handle each directory recursively.

Useful options:

-s	Summarize (give total for each argument)
-h	Display sizes in 'human readable format', using units of: B=bytes, K=kilobytes, M=megabytes, G=gigabytes
--exclude <pattern>	Exclude files or directories which match the given pattern

Examples:

Give total disk usage for the current directory and all subdirectories

du -s

Give total disk usage for each file and each directory (and their subdirectories). Print the sizes in units of bytes, kilobytes, megabytes, gigabytes.

du -sh *

Give total disk usage for each file and subdirectory of the home directory, excluding the Backup subdirectory. With our AFS filesystem at CABI, the Backup subdirectory of the home directory is a snapshot of the home directory, updated every night, so it is not meaningful to include Backup in the disk usage summary.

du -s --exclude=Backup ~/*

ping [options] <computer>

Send a signal across the network to another computer. Useful for testing whether that computer is running and accessible on the network, or if your Linux computer can reach the network. You can specify the computer by name or by IP address. If the computer that you are pinging is accessible via the network from your computer, you will get a response every second until you exit. Exit the program with control-c.

Examples:

See if the computer 'rimmer' is accessible via the network.

ping rimmer

bc [options] [files]

Arbitrary-precision calculator.
Very useful when you need to calculate something but are too proud or lazy to go get a calculator. In interactive mode you can calculate expressions from command-line input. To quit interactive mode, press control-d or type 'quit'. Input can be read from files. It can also be used in bash (in a normal shell or in a script) to calculate expressions. Actually bc is a whole calculator language. For details on programming in bc (for example for defining your own functions, or for using loops) see the manpage.

Useful option:

-l	Define the standard math library. Needed for the functions listed below.

(The standard math library is not needed for sqrt, but you should use it to avoid low precision.)

Some useful functions:

sqrt( )	Square root
s( )	Sine (argument in radians)
c( )	Cosine (argument in radians)
a( )	Arctangent (result in radians)
l( )	Natural logarithm
e( )	Exponential

Examples:

Start bc in interactive mode with the standard math library defined, and calculate the sine of 10 degrees. Makes use of the fact that arctan(1)=PI/4 (=45 degrees).

bc -l
pi = 4*a(1)
pi
3.14159265358979323844
degs2rads = pi/180
s(10*degs2rads)
.17364817766693034884

Define a function for calculating the tangent of an angle given in radians

bc -l
pi = 4*a(1)
define tan_degrees (x) {
   auto y
   y = x * pi/180.
   return (s(y) / c(y));
}
tan_degrees(117)
-1.96261050550515058239

Start bc in interactive mode and calculate 7.5*7.5*7.5

bc
7.5^3
421.875

In a bash shell (or script), calculate the 2nd-order Bessel function of 3.5

x=`echo "j(2, 3.5)" | bc -l`
echo $x
.45862918419430748350

units [options] [from-units [to-units]]

Convert between units. If the units are not specified, the program enters interactive mode; exit by typing ^d (control-d)

Examples:

You want to know how many ml are in 3 teaspoons and how many furlongs are in a marathon

units
2112 units, 59 prefixes

You have: 3 teaspoons
You want: ml

You have: marathon
You want: furlong

An English chap with whom you work tells you that the magnet can go another fortnight without a helium refill, and you want to know about what he's talking

units fortnights weeks
* 2
/ 0.5

This page was last updated on 1/28/2008.