Arguments: $1, $#, $@ and shift

Everything you type after a script's name arrives inside it as positional parameters: $1 is the first word, $2 the second, and so on up to $9; from ten on you need braces, ${10}. $0 is the name the script was invoked with, useful in error messages. $# is how many arguments there are, and $@ is all of them. These are the same variables you met inside functions in the previous lesson, now at script level.

~% cat lab/l88/args.sh
#!/bin/sh
echo "script: $0"
echo "count:  $#"
echo "first:  $1"
echo "all:    $@"
~% lab/l88/args.sh red "light blue" green
script: lab/l88/args.sh
count:  3
first:  red
all:    red light blue green

Checking what you were given

A script that needs a file name should refuse to run without one, and say how to use it. The convention is a usage line on standard error and a non-zero exit, so that the caller, a human or another script, knows something went wrong. Use $# for the count and -z "$1" when an empty argument is also unacceptable.

#!/bin/sh
if [ $# -lt 2 ]; then
    echo "usage: $0 SOURCE DEST" >&2
    exit 1
fi
src="$1"
dst="$2"
cp "$src" "$dst" && echo "copied $src to $dst"

$@ versus $*

Both expand to all the arguments, and unquoted they behave identically, which is to say badly: words get split again. The difference appears in double quotes. "$*" is one string with the arguments joined by spaces. "$@" is each argument as its own word, spaces inside preserved. When you pass arguments on to another command or loop over them, "$@" is almost always what you mean; "$*" is for building a single message.

~% cat lab/l88/each.sh
#!/bin/sh
for a in "$@"; do echo "arg: $a"; done
echo "---"
for a in $*; do echo "word: $a"; done
~% lab/l88/each.sh "light blue" green
arg: light blue
arg: green
---
word: light
word: blue
word: green

shift

shift throws away $1 and renumbers the rest: the old $2 becomes $1, $# drops by one. It is how you process arguments one at a time, or peel off options before the real arguments. shift 2 drops two at once. A loop while [ $# -gt 0 ]; do … shift; done visits every argument and also lets you react to flags like -v. For anything more elaborate the getopts builtin exists, but the shift loop covers most small tools.

#!/bin/sh
verbose=0
if [ "$1" = "-v" ]; then
    verbose=1
    shift
fi
echo "$# files to process: $*"
[ "$verbose" -eq 1 ] && echo "(verbose mode)"