Classic Shell Scripting

The case Statement

If you need to check a variable for one of many values, you could use a cascading series of if and elif tests, together with test:

if [ "X$1" = "X-f" ]
then
    ...     Code for -f option
elif [ "X$1" = "X-d" ] || [ "X$1" = "X--directory" ]  # long option allowed
then
    ...     Code for -d option
else
    echo $1: unknown option >&2
    exit 1
fi

However, this is awkward to write and difficult to read. (The >&2 in the echo command sends the output to standard error. This is described in Section 7.3.2.) Instead, the shell's case construct should be used for pattern matching:

case $1 in
-f)
    ...     Code for -f option
    ;;
-d | --directory)  # long option allowed
    ...     Code for -d option
    ;;
*)
    echo $1: unknown option >&2
    exit 1
    # ;; is good form before `esac', but not required
esac

As can be seen, the value to be tested appears between case and in. Double-quoting the value, while not necessary, doesn't hurt either. The value is tested against each list of shell patterns in turn. When one matches, the corresponding body of code, up to the ;;, is executed. Multiple patterns may be used, separated by the | character, which in this context means "or." The patterns may contain any shell wildcard characters, and variable, command, and arithmetic substitutions are performed on the value before it is used for pattern matching.

The unbalanced right parenthesis after each pattern list is perhaps surprising; this is the only instance in the shell language of unbalanced delimiters. (In Section 14.3.7, we will see ...