Function Declaration

A Bash function is a named, reusable block of commands that executes within the current shell execution environment. It acts as a user-defined command that supersedes external executables and shell built-ins of the same name during the shell’s command resolution process.

Bash supports three distinct syntactic forms for declaring functions. In all forms, the body of the function can be any valid Bash compound command, optionally followed by standard shell redirections.

1. POSIX Standard Syntax (Recommended) This is the most portable declaration method, compatible with strictly POSIX-compliant shells.

function_name() compound-command [redirection]

2. Bash-Specific Keyword Syntax This form uses the function reserved word and omits the parentheses. It is specific to Bash and KornShell (ksh).

function function_name compound-command [redirection]

3. Combined Syntax This form uses both the function keyword and parentheses. While valid in Bash, it is redundant and generally discouraged in strict style guides.

function function_name() compound-command [redirection]

Compound Commands as Function Bodies

While the grouping command { ... } is the most common function body, Bash allows a function body to be any compound command. This includes subshells ( ... ), arithmetic evaluations (( ... )), conditional expressions [[ ... ]], and loops (e.g., for, while, case).

# Grouping command (executes in current shell environment)
my_func() { echo "Standard block"; }


# Subshell (executes in a child process)
my_func() ( cd /tmp && ls )


# Arithmetic evaluation
my_func() (( var++ ))


# Conditional expression
my_func() [[ -f "$1" ]]


# Loop
my_func() for i in "$@"; do echo "$i"; done

Lexical and Syntactic Rules

• Reserved Word Delimiting: The opening curly brace { is a reserved word, not a shell metacharacter. It must be delimited by whitespace or a shell metacharacter. If the first command inside the block starts with a metacharacter (such as (, <, or >), no space is required after the {. Conversely, ( and ) are metacharacters, so function_name(){ is valid without a space before the brace.
• Command Termination: When using the { ... } grouping command, the list of commands within the body must be explicitly terminated. The final command before the closing brace } must end with a newline, a semicolon (;), or an ampersand (&). Subshells ( ... ) do not require this explicit termination before the closing parenthesis.

# Valid: Space after {, terminated with ;
my_func() { command1; command2; }


# Valid: No space after { because ( and > are metacharacters
my_func(){(command1);}
my_func(){>out.txt command1;}


# Invalid: Missing command terminator before }
my_func() { command1; command2 }

Declaration Mechanics

• Sequential Parsing: Bash is an interpreted language that parses sequentially. A function must be declared lexically prior to its invocation. The shell does not hoist function declarations.
• Symbol Table Overwriting: If a function is declared with a name that already exists in the shell’s function environment, the new declaration overwrites the previous one. Exception: If the existing function has been marked as read-only using readonly -f, Bash will throw an error (bash: function_name: readonly function) and refuse to overwrite it.
• Command Precedence: Once declared, functions take high precedence in the shell’s execution hierarchy. In default Bash, the resolution order is: Aliases > Functions > Built-ins (both special and regular) > External Executables (in $PATH). Note: Special built-ins only supersede functions when Bash is invoked in strict POSIX mode.
• Removal: A declared function remains in the current shell’s memory until the shell terminates or the function is explicitly removed from the symbol table using the unset built-in with the -f flag.

unset -f function_name

Master Bash with Deep Grasping Methodology!

Learn More