📖 For those who wanna learn Bash
Latest commit 81939f8 Dec 25, 2016 @denysdovhan 2.9.0
Failed to load latest commit information.
bin Add -h alias. May 26, 2016
translations OS X → macOS Aug 12, 2016
.editorconfig Add .editorconfig. Dec 31, 2015
.gitignore Add ignore files. Dec 31, 2015
.npmignore Add ignore files. Dec 31, 2015
README.md OS X → macOS Aug 12, 2016
package.json 2.9.0 Dec 25, 2016


bash-handbook CC 4.0

NPM version Gitter

This document was written for those who want to learn Bash without diving in too deeply.

Tip: Try learnyoubash — an interactive workshopper based on this handbook!

Node Packaged Manuscript

You can install this handbook using npm. Just run:

$ npm install -g bash-handbook

You should be able to run bash-handbook at the command line now. This will open the manual in your selected $PAGER. Otherwise, you may continue reading on here.

The source is available here: https://github.com/denysdovhan/bash-handbook


Currently, there are these translations of bash-handbook:

Request another translation

Table of Contents


If you are a developer, then you know the value of time. Optimizing your work process is one of the most important aspects of the job.

In that path towards efficiency and productivity, we are often posed with actions that must be repeated over and over again, like:

  • taking a screenshot and uploading it to a server
  • processing text that may come in many shapes and forms
  • converting files between different formats
  • parsing a program's output

Enter Bash, our savior.

Bash is a Unix shell written by Brian Fox for the GNU Project as a free software replacement for the Bourne shell. It was released in 1989 and has been distributed as the Linux and macOS default shell for a long time.

So why do we need to learn something that was written more than 30 years ago? The answer is simple: this something is today one of the most powerful and portable tools for writing efficient scripts for all Unix-based systems. And that's why you should learn bash. Period.

In this handbook, I'm going to describe the most important concepts in bash with examples. I hope this compendium will be helpful to you.

Shells and modes

The user bash shell can work in two modes - interactive and non-interactive.

Interactive mode

If you are working on Ubuntu, you have seven virtual terminals available to you. The desktop environment takes place in the seventh virtual terminal, so you can return to a friendly GUI using the Ctrl-Alt-F7 keybinding.

You can open the shell using the Ctrl-Alt-F1 keybinding. After that, the familiar GUI will disappear and one of the virtual terminals will be shown.

If you see something like this, then you are working in interactive mode:


Here you can enter a variety of Unix commands, such as ls, grep, cd, mkdir, rm and see the result of their execution.

We call this shell interactive because it interacts directly with the user.

Using a virtual terminal is not really convenient. For example, if you want to edit a document and execute another command at the same time, you are better off using virtual terminal emulators like:

Non-interactive mode

In non-interactive mode, the shell reads commands from a file or a pipe and executes them. When the interpreter reaches the end of the file, the shell process terminates the session and returns to the parent process.

Use the following commands for running the shell in non-interactive mode:

sh /path/to/script.sh
bash /path/to/script.sh

In the example above, script.sh is just a regular text file that consists of commands the shell interpreter can evaluate and sh or bash is the shell's interpreter program. You can create script.sh using your preferred text editor (e.g. vim, nano, Sublime Text, Atom, etc).

You can also simplify invoking the script by making it an executable file using the chmod command:

chmod +x /path/to/script.sh

Additionally, the first line in the script must indicate which program it should use to run the file, like so:

echo "Hello, world!"

Or if you prefer to use sh instead of bash, change #!/bin/bash to #!/bin/sh. This #! character sequence is known as the shebang. Now you can run the script like this:


A handy trick we used above is using echo to print text to the terminal screen.

Another way to use the shebang line is as follows:

#!/usr/bin/env bash
echo "Hello, world!"

The advantage of this shebang line is it will search for the program (in this case bash) based on the PATH environment variable. This is often preferred over the first method shown above, as the location of a program on a filesystem cannot always be assumed. This is also useful if the PATH variable on a system has been configured to point to an alternate version of the program. For instance, one might install a newer version of bash while preserving the original version and insert the location of the newer version into the PATH variable. The use of #!/bin/bash would result in using the original bash, while #!/usr/bin/env bash would make use of the newer version.

Exit codes

Every command returns an exit code (return status or exit status). A successful command always returns 0 (zero-code), and a command that has failed returns a non-zero value (error code). Failure codes must be positive integers between 1 and 255.

Another handy command we can use when writing a script is exit. This command is used to terminate the current execution and deliver an exit code to the shell. Running an exit code without any arguments, will terminate the running script and return the exit code of the last command executed before exit.

When a program terminates, the shell assigns its exit code to the $? environment variable. The $? variable is how we usually test whether a script has succeeded or not in its execution.

In the same way we can use exit to terminate a script, we can use the return command to exit a function and return an exit code to the caller. You can use exit inside a function too and this will exit the function and terminate the program.


Scripts may contain comments. Comments are special statements ignored by the shell interpreter. They begin with a # symbol and continue on to the end of the line.

For example:

# This script will print your username.

Tip: Use comments to explain what your script does and why.


Like in most programming languages, you can also create variables in bash.

Bash knows no data types. Variables can contain only numbers or a string of one or more characters. There are three kinds of variables you can create: local variables, environment variables and variables as positional arguments.

Local variables

Local variables are variables that exist only within a single script. They are inaccessible to other programs and scripts.

A local variable can be declared using = sign (as a rule, there should not be any spaces between a variable's name, = and its value) and its value can be retrieved using the $ sign. For example:

username="denysdovhan"  # declare variable
echo $username          # display value
unset username          # delete variable

We can also declare a variable local to a single function using the local keyword. Doing so causes the variable to disappear when the function exits.

local local_var="I'm a local value"

Environment variables

Environment variables are variables accessible to any program or script running in current shell session. They are created just like local variables, but using the keyword export instead.

export GLOBAL_VAR="I'm a global variable"

There are a lot of global variables in bash. You will meet these variables fairly often, so here is a quick lookup table with the most practical ones:

Variable Description
$HOME The current user's home directory.
$PATH A colon-separated list of directories in which the shell looks for commands.
$PWD The current working directory.
$RANDOM Random integer between 0 and 32767.
$UID The numeric, real user ID of the current user.
$PS1 The primary prompt string.
$PS2 The secondary prompt string.

Follow this link to see an extended list of environment variables in Bash.

Positional parameters

Positional parameters are variables allocated when a function is evaluated and are given positionally. The following table lists positional parameter variables and other special variables and their meanings when you are inside a function.

Parameter Description
$0 Script's name.
$1 … $9 The parameter list elements from 1 to 9.
${10} … ${N} The parameter list elements from 10 to N.
$* or $@ All positional parameters except $0.
$# The number of parameters, not counting $0.
$FUNCNAME The function name (has a value only inside a function).

In the example below, the positional parameters will be $0='./script.sh', $1='foo' and $2='bar':

./script.sh foo bar

Variables may also have default values. We can define as such using the following syntax:

 # if variables are empty, assign them default values
: ${VAR:='default'}
: ${$1:='first'}
# or

Shell expansions

Expansions are performed on the command line after it has been split into tokens. In other words, these expansions are a mechanism to calculate arithmetical operations, to save results of commands' executions and so on.

If you are interested, you can read more about shell expansions.

Brace expansion

Brace expansion allows us to generate arbitrary strings. It's similar to filename expansion. For example:

echo beg{i,a,u}n # begin began begun

Also brace expansions may be used for creating ranges, which are iterated over in loops.

echo {0..5} # 0 1 2 3 4 5
echo {00..8..2} # 00 02 04 06 08

Command substitution

Command substitution allow us to evaluate a command and substitute its value into another command or variable assignment. Command substitution is performed when a command is enclosed by `` or $(). For example, we can use it as follows:

now=`date +%T`
# or
now=$(date +%T)

echo $now # 19:08:26

Arithmetic expansion

In bash we are free to do any arithmetical operations. But the expression must enclosed by $(( )) The format for arithmetic expansions is:

result=$(( ((10 + 5*3) - 7) / 2 ))
echo $result # 9

Within arithmetic expansions, variables should generally be used without a $ prefix:

echo $(( x + y ))     # 11
echo $(( ++x + y++ )) # 12
echo $(( x + y ))     # 13

Double and single quotes

There is an important difference between double and single quotes. Inside double quotes variables or command substitutions are expanded. Inside single quotes they are not. For example:

echo "Your home: $HOME" # Your home: /Users/<username>
echo 'Your home: $HOME' # Your home: $HOME

Take care to expand local variables and environment variables within quotes if they could contain whitespace. As an innocuous example, consider using echo to print some user input:

INPUT="A string  with   strange    whitespace."
echo $INPUT   # A string with strange whitespace.
echo "$INPUT" # A string  with   strange    whitespace.

The first echo is invoked with 5 separate arguments — $INPUT is split into separate words, echo prints a single space character between each. In the second case, echo is invoked with a single argument (the entire $INPUT value, including whitespace).

Now consider a more serious example:

FILE="Favorite Things.txt"
cat $FILE   # attempts to print 2 files: `Favorite` and `Things.txt`
cat "$FILE" # prints 1 file: `Favorite Things.txt`

While the issue in this example could be resolved by renaming FILE to Favorite-Things.txt, consider input coming from an environment variable, a positional parameter, or the output of another command (find, cat, etc). If the input might contain whitespace, take care to wrap the expansion in quotes.


Like in other programming languages, an array in bash is a variable that allows you to refer to multiple values. In bash, arrays are also zero-based, that is, the first element in an array has index 0.

When dealing with arrays, we should be aware of the special environment variable IFS. IFS, or Input Field Separator, is the character that separates elements in an array. The default value is an empty space IFS=' '.

Array declaration

In bash you create an array by simply assigning a value to an index in the array variable:


Array variables can also be created using compound assignments such as:

fruits=(Apple Pear Plum)

Array expansion

Individual array elements are expanded similar to other variables:

echo ${fruits[1]} # Pear

The entire array can be expanded by using * or @ in place of the numeric index:

echo ${fruits[*]} # Apple Pear Plum
echo ${fruits[@]} # Apple Pear Plum

There is an important (and subtle) difference between the two lines above: consider an array element containing whitespace:

fruits[1]="Desert fig"

We want to print each element of the array on a separate line, so we try to use the printf builtin:

printf "+ %s\n" ${fruits[*]}
# + Apple
# + Desert
# + fig
# + Plum

Why were Desert and fig printed on separate lines? Let's try to use quoting:

printf "+ %s\n" "${fruits[*]}"
# + Apple Desert fig Plum

Now everything is on one line — that's not what we wanted! Here's where ${fruits[@]} comes into play:

printf "+ %s\n" "${fruits[@]}"
# + Apple
# + Desert fig
# + Plum

Within double quotes, ${fruits[@]} expands to a separate argument for each element in the array; whitespace in the array elements is preserved.

Array slice

Besides, we can extract a slice of array using the slice operators:

echo ${fruits[@]:0:2} # Apple Desert fig

In the example above, ${fruits[@]} expands to the entire contents of the array, and :0:2 extracts the slice of length 2, that starts at index 0.

Adding elements into an array

Adding elements into an array is quite simple too. Compound assignments are specially useful in this case. We can use them like this:

fruits=(Orange "${fruits[@]}" Banana Cherry)
echo ${fruits[@]} # Orange Apple Desert fig Plum Banana Cherry

The example above, ${fruits[@]} expands to the entire contents of the array and substitutes it into the compound assignment, then assigns the new value into the fruits array mutating its original value.

Deleting elements from an array

To delete an element from an array, use the unset command:

unset fruits[0]
echo ${fruits[@]} # Apple Desert fig Plum Banana Cherry

Streams, pipes and lists

Bash has powerful tools for working with other programs and their outputs. Using streams we can send the output of a program into another program or file and thereby write logs or whatever we want.

Pipes give us opportunity to create conveyors and control the execution of commands.

It is paramount we understand how to use this powerful and sophisticated tool.


Bash receives input and sends output as sequences or streams of characters. These streams may be redirected into files or one into another.

There are three descriptors:

Code Descriptor Description
0 stdin The standard input.
1 stdout The standard output.
2 stderr The errors output.

Redirection makes it possible to control where the output of a command goes to, and where the input of a command comes from. For redirecting streams these operators are used:

Operator Description
> Redirecting output
&> Redirecting output and error output
&>> Appending redirected output and error output
< Redirecting input
<< Here documents syntax
<<< Here strings

Here are few examples of using redirections:

# output of ls will be written to list.txt
ls -l > list.txt

# append output to list.txt
ls -a >> list.txt

# all errors will be written to errors.txt
grep da * 2> errors.txt

# read from errors.txt
less < errors.txt


We could redirect standard streams not only in files, but also to other programs. Pipes let us use the output of a program as the input of another.

In the example below, command1 sends its output to command2, which then passes it on to the input of command3:

command1 | command2 | command3

Constructions like this are called pipelines.

In practice, this can be used to process data through several programs. For example, here the output of ls -l is sent to the grep program, which prints only files with a .md extension, and this output is finally sent to the less program:

ls -l | grep .md$ | less

The exit status of a pipeline is normally the exit status of the last command in the pipeline. The shell will not return a status until all the commands in the pipeline have completed. If you want your pipelines to be considered a failure if any of the commands in the pipeline fail, you should set the pipefail option with:

set -o pipefail

Lists of commands

A list of commands is a sequence of one or more pipelines separated by ;, &, && or || operator.

If a command is terminated by the control operator &, the shell executes the command asynchronously in a subshell. In other words, this command will be executed in the background.

Commands separated by a ; are executed sequentially: one after another. The shell waits for the finish of each command.

# command2 will be executed after command1
command1 ; command2

# which is the same as

Lists separated by && and || are called AND and OR lists, respectively.

The AND-list looks like this:

# command2 will be executed if, and only if, command1 finishes successfully (returns 0 exit status)
command1 && command2

The OR-list has the form:

# command2 will be executed if, and only if, command1 finishes unsuccessfully (returns code of error)
command1 || command2

The return code of an AND or OR list is the exit status of the last executed command.

Conditional statements

Like in other languages, Bash conditionals let us decide to perform an action or not. The result is determined by evaluating an expression, which should be enclosed in [[ ]].

Conditional expression may contain && and || operators, which are AND and OR accordingly. Besides this, there many other handy expressions.

There are two different conditional statements: if statement and case statement.

Primary and combining expressions

Expressions enclosed inside [[ ]] (or [ ] for sh) are called test commands or primaries. These expressions help us to indicate results of a conditional. In the tables below, we are using [ ], because it works for sh too. Here is an answer about the difference between double and single square brackets in bash.

Working with the file system:

Primary Meaning
[ -e FILE ] True if FILE exists.
[ -f FILE ] True if FILE exists and is a regular file.
[ -d FILE ] True if FILE exists and is a directory.
[ -s FILE ] True if FILE exists and not empty (size more than 0).
[ -r FILE ] True if FILE exists and is readable.
[ -w FILE ] True if FILE exists and is writable.
[ -x FILE ] True if FILE exists and is executable.
[ -L FILE ] True if FILE exists and is symbolic link.
[ FILE1 -nt FILE2 ] FILE1 is newer than FILE2.
[ FILE1 -ot FILE2 ] FILE1 is older than FILE2.

Working with strings:

Primary Meaning
[ -z STR ] STR is empty (the length is zero).
[ -n STR ] STR is not empty (the length is non-zero).
[ STR1 == STR2 ] STR1 and STR2 are equal.
[ STR1 != STR2 ] STR1 and STR2 are not equal.

Arithmetic binary operators:

Primary Meaning
[ ARG1 -eq ARG2 ] ARG1 is equal to ARG2.
[ ARG1 -ne ARG2 ] ARG1 is not equal to ARG2.
[ ARG1 -lt ARG2 ] ARG1 is less than ARG2.
[ ARG1 -le ARG2 ] ARG1 is less than or equal to ARG2.
[ ARG1 -gt ARG2 ] ARG1 is greater than ARG2.
[ ARG1 -ge ARG2 ] ARG1 is greater than or equal to ARG2.

Conditions may be combined using these combining expressions:

Operation Effect
[ ! EXPR ] True if EXPR is false.
[ (EXPR) ] Returns the value of EXPR.
[ EXPR1 -a EXPR2 ] Logical AND. True if EXPR1 and EXPR2 are true.
[ EXPR1 -o EXPR2 ] Logical OR. True if EXPR1 or EXPR2 are true.

Sure, there are more useful primaries and you can easily find them in the Bash man pages.

Using an if statement

if statements work the same as in other programming languages. If the expression within the braces is true, the code between then and fi is executed. fi indicates the end of the conditionally executed code.

# Single-line
if [[ 1 -eq 1 ]]; then echo "true"; fi

# Multi-line
if [[ 1 -eq 1 ]]; then
  echo "true"

Likewise, we could use an if..else statement such as:

# Single-line
if [[ 2 -ne 1 ]]; then echo "true"; else echo "false"; fi

# Multi-line
if [[ 2 -ne 1 ]]; then
  echo "true"
  echo "false"

Sometimes if..else statements are not enough to do what we want to do. In this case we shouldn't forget about the existence of if..elif..else statements, which always come in handy.

Look at the example below:

if [[ `uname` == "Adam" ]]; then
  echo "Do not eat an apple!"
elif [[ `uname` == "Eva" ]]; then
  echo "Do not take an apple!"
  echo "Apples are delicious!"

Using a case statement

If you are confronted with a couple of different possible actions to take, then using a case statement may be more useful than nested if statements. For more complex conditions use case like below:

case "$extension" in
    echo "It's image with jpeg extension."
    echo "It's image with png extension."
    echo "Oh, it's a giphy!"
    echo "Woops! It's not image!"

Each case is an expression matching a pattern. The | sign is used for separating multiple patterns, and the ) operator terminates a pattern list. The commands for the first match are executed. * is the pattern for anything else that doesn't match the defined patterns. Each block of commands should be divided with the ;; operator.


Here we won't be surprised. As in any programming language, a loop in bash is a block of code that iterates as long as the control conditional is true.

There are four types of loops in Bash: for, while, until and select.

for loop

The for is very similar to its sibling in C. It looks like this:

for arg in elem1 elem2 ... elemN
  # statements

During each pass through the loop, arg takes on the value from elem1 to elemN. Values may also be wildcards or brace expansions.

Also, we can write for loop in one line, but in this case there needs to be a semicolon before do, like below:

for i in {1..5}; do echo $i; done

By the way, if for..in..do seems a little bit weird to you, you can also write for in C-like style such as:

for (( i = 0; i < 10; i++ )); do
  echo $i

for is handy when we want to do the same operation over each file in a directory. For example, if we need to move all .bash files into the script folder and then give them execute permissions, our script would look like this:


for FILE in $HOME/*.bash; do
  mv "$FILE" "${HOME}/scripts"
  chmod +x "${HOME}/scripts/${FILE}"

while loop

The while loop tests a condition and loops over a sequence of commands so long as that condition is true. A condition is nothing more than a primary as used in if..then conditions. So a while loop looks like this:

while [[ condition ]]
  # statements

Just like in the case of the for loop, if we want to write do and condition in the same line, then we must use a semicolon before do.

A working example might look like this:


# Squares of numbers from 0 through 9
while [[ $x -lt 10 ]]; do # value of x is less than 10
  echo $(( x * x ))
  x=$(( x + 1 )) # increase x

until loop

The until loop is the exact opposite of the while loop. Like a while it checks a test condition, but it keeps looping as long as this condition is false:

until [[ condition ]]; do

select loop

The select loop helps us to organize a user menu. It has almost the same syntax as the for loop:

select answer in elem1 elem2 ... elemN
  # statements

The select prints all elem1..elemN on the screen with their sequence numbers, after that it prompts the user. Usually it looks like $? (PS3 variable). The answer will be saved in answer. If answer is the number between 1..N, then statements will execute and select will go to the next iteration — that's because we should use the break statement.

A working example might look like this:


PS3="Choose the package manager: "
select ITEM in bower npm gem pip
  echo -n "Enter the package name: " && read PACKAGE
  case $ITEM in
    bower) bower install $PACKAGE ;;
    npm)   npm   install $PACKAGE ;;
    gem)   gem   install $PACKAGE ;;
    pip)   pip   install $PACKAGE ;;
  break # avoid infinite loop

This example, asks the user what package manager {s,he} would like to use. Then, it will ask what package we want to install and finally proceed to install it.

If we run this, we will get:

$ ./my_script
1) bower
2) npm
3) gem
4) pip
Choose the package manager: 2
Enter the package name: bash-handbook
<installing bash-handbook>

Loop control

There are situations when we need to stop a loop before its normal ending or step over an iteration. In these cases, we can use the shell built-in break and continue statements. Both of these work with every kind of loop.

The break statement is used to exit the current loop before its ending. We have already met with it.

The continue statement steps over one iteration. We can use it as such:

for (( i = 0; i < 10; i++ )); do
  if [[ $(( i % 2 )) -eq 0 ]]; then continue; fi
  echo $i

If we run the example above, it will print all odd numbers from 0 through 9.


In scripts we have the ability to define and call functions. As in any programming language, functions in bash are chunks of code, but there are differences.

In bash, functions are a sequence of commands grouped under a single name, that is the name of the function. Calling a function is the same as calling any other program, you just write the name and the function will be invoked.

We can declare our own function this way:

my_func () {
  # statements

my_func # call my_func

We must declare functions before we can invoke them.

Functions can take on arguments and return a result — exit code. Arguments, within functions, are treated in the same manner as arguments given to the script in non-interactive mode — using positional parameters. A result code can be returned using the return command.

Below is a function that takes a name and returns 0, indicating successful execution.

# function with params
greeting () {
  if [[ -n $1 ]]; then
    echo "Hello, $1!"
    echo "Hello, unknown!"
  return 0

greeting Denys  # Hello, Denys!
greeting        # Hello, unknown!

We already discussed exit codes. The return command without any arguments returns the exit code of the last executed command. Above, return 0 will return a successful exit code. 0.


The shell gives us tools for debugging scripts. If we want to run a script in debug mode, we use a special option in our script's shebang:

#!/bin/bash options

These options are settings that change shell behavior. The following table is a list of options which might be useful to you:

Short Name Description
-f noglob Disable filename expansion (globbing).
-i interactive Script runs in interactive mode.
-n noexec Read commands, but don't execute them (syntax check).
pipefail Make pipelines fail if any commands fail, not just if the final command fail.
-t Exit after first command.
-v verbose Print each command to stderr before executing it.
-x xtrace Print each command and its expanded arguments to stderr before executing it.

For example, we have script with -x option such as:

#!/bin/bash -x

for (( i = 0; i < 3; i++ )); do
  echo $i

This will print the value of the variables to stdout along with other useful information:

$ ./my_script
+ (( i = 0 ))
+ (( i < 3 ))
+ echo 0
+ (( i++  ))
+ (( i < 3 ))
+ echo 1
+ (( i++  ))
+ (( i < 3 ))
+ echo 2
+ (( i++  ))
+ (( i < 3 ))

Sometimes we need to debug a part of a script. In this case using the set command is convenient. This command can enable and disable options. Options are turned on using - and turned off using +:


echo "xtrace is turned off"
set -x
echo "xtrace is enabled"
set +x
echo "xtrace is turned off again"


I hope this small handbook was interesting and helpful. To be honest, I wrote this handbook for myself so as to not forget the bash basics. I tried to write concisely but meaningfully, and I hope you will appreciate that.

This handbook narrates my own experience with Bash. It does not purport to be comprehensive, so if you still want more, please run man bash and start there.

Contributions are absolutely welcome and I will be grateful for any corrections or questions you can send my way. For all of that create a new issue.

Thanks for reading this handbook!

Want to learn more?

Here's a list of other literature covering Bash:

Other resources

  • awesome-bash is a curated list of Bash scripts and resources
  • awesome-shell is another curated list of shell resources
  • bash-it provides a solid framework for using, developing and maintaining shell scripts and custom commands for your daily work.
  • dotfiles.github.io is a good source of pointers to the various dotfiles collections and shell frameworks available for bash and other shells.
  • learnyoubash helps you write your first bash script
  • shellcheck is a static analysis tool for shell scripts. You can either use it from a web page at www.shellcheck.net or run it from the command line. Installation instructions are on the koalaman/shellcheck github repository page.

Finally, Stack Overflow has many questions that are tagged as bash that you can learn from and is a good place to ask if you're stuck.


CC 4.0 © Denys Dovhan