Shell by Example: Command Line Arguments POSIX 

#!/bin/sh

Shell scripts can access command-line arguments through special variables. This example demonstrates how to work with positional parameters.

Set some arguments for demonstration:

 

set -- hello world foo bar

$0 is the script name (or shell name in interactive mode):

 

echo "Script name: $0"

\(1, \)2, etc. are positional parameters:

 

echo "First argument: $1"
echo "Second argument: $2"
echo "Third argument: $3"

$# is the number of arguments:

 

echo "Number of arguments: $#"

$@ expands to all arguments as separate words:

 

echo "All arguments (\$@):"
for arg in "$@"; do
  echo "  - $arg"
done

$* expands to all arguments as a single word when quoted:

 

echo "All arguments (\$*):"
echo "  $*"

The difference matters with spaces in arguments:

 

set -- "hello world" "foo bar"
echo "With quoted \$@:"
for arg in "$@"; do
  echo "  [$arg]"
done

echo "With quoted \$*:"
for arg in "$*"; do
  echo "  [$arg]"
done

Always use “$@” (quoted) to preserve argument boundaries.

shift removes the first argument and shifts others down:

 

set -- first second third fourth
echo "Before shift: $1 $2 $3 $4"
shift
echo "After shift: $1 $2 $3"
shift 2
echo "After shift 2: $1"

Common pattern: Process all arguments with shift:

 

set -- -v --name=test file1.txt file2.txt

echo "Processing arguments with shift:"
while [ $# -gt 0 ]; do
  echo "  Processing: $1"
  shift
done

Check if arguments were provided:

 

check_args() {
  if [ $# -eq 0 ]; then
    echo "No arguments provided"
    return 1
  fi
  echo "Got $# argument(s)"
}

check_args
check_args one two

Default values for missing arguments:

 

set -- customvalue

name="${1:-default_name}"
port="${2:-8080}"
echo "Name: $name, Port: $port"

\({var:-default} uses default if var is unset or empty \){var:=default} also assigns the default to var

Require an argument (exit if missing):

 

require_arg() {
  : "${1:?Error: First argument required}"
  echo "Got required arg: $1"
}

require_arg "provided"
# require_arg  # Would cause an error

Arguments beyond $9 need braces:

 

set -- a b c d e f g h i j k l

echo "Arg 1: $1"
echo "Arg 9: $9"
echo "Arg 10: ${10}"
echo "Arg 12: ${12}"

Check for specific number of arguments:

 

exactly_two() {
  if [ $# -ne 2 ]; then
    echo "Usage: exactly_two <arg1> <arg2>"
    return 1
  fi
  echo "Got exactly two: $1 and $2"
}

exactly_two one two
exactly_two one

Parse arguments into variables:

 

set -- source.txt destination.txt --verbose

src="$1"
dst="$2"
verbose=""

for arg in "$@"; do
  case "$arg" in
    --verbose | -v) verbose=true ;;
  esac
done

echo "Source: $src"
echo "Destination: $dst"
echo "Verbose: ${verbose:-false}"

$$ is the script’s process ID:

 

echo "Script PID: $$"

$! is the PID of the last background command:

 

sleep 1 &
echo "Background PID: $!"
wait
Bash

Bash provides additional variables:

  • $BASHPID - PID of current bash process (differs in subshells)
  • ${@:n:m} - Slice of arguments starting at n, m elements
  • ${!#} - The last argument
 

echo "Last argument: ${!#}"
echo "Args 2-3: ${@:2:2}"

Common argument handling pattern:

 

handle_args() {
  while [ $# -gt 0 ]; do
    case "$1" in
      -h | --help)
        echo "Usage: script [options] files..."
        return 0
        ;;
      -v | --verbose)
        echo "Verbose mode on"
        shift
        ;;
      --)
        shift
        break # Rest are positional args
        ;;
      -*)
        echo "Unknown option: $1" >&2
        return 1
        ;;
      *)
        break # Start of positional args
        ;;
    esac
  done

  echo "Remaining arguments: $*"
}

handle_args -v -- file1 file2

« Traps | Getopts »

by Jose Diaz-Gonzalez | inspired by Go by Example | source