The shell's case statement (44.5) has some advantages over the test command (44.20)- for instance, case can do pattern matching. But test has the -a and -o "and" and "or" operators; those don't seem easy to do with case. And test isn't built in to some older shells, so using case may be faster.
Here's a way to test two things with one case statement. It won't solve all your problems. If you think carefully about the possible values the variables you're testing can have, though, this might do the trick. Use a separator (delimiter) character between the two variables.
In the example below, I've picked a slash (/
).
You could use almost any character that isn't used in
case pattern matching (44.6)
and that won't be stored in either $#
or $1
.
The case below tests the command-line arguments of a script:
case "$#/$1" in 1/-f) redodb=yes ;; 0/) ;; *) echo "Usage: $myname [-f]" 1>&2; exit 1 ;; esac
If there's one argument ($#
is 1
) and the argument
($1
) is exactly -f
, the first pattern matches, and the
redodb variable is set.
If there's no argument, $#
will be 0
and $1
will
be empty, so the second pattern matches.
Otherwise, something is wrong; the third pattern matches, the script
prints an error and exits.
Of course, you can do a lot more this way than just testing command-line arguments.
-