We have regular expressions, and they have metacharacters and character classes.

if ( $char =~ /\d/ ) {
    print "It's a digit.\n";
}
if ( $char =~ /\w/ ) {
    print "It's a word-like character.\n";
}
if ( $char =~ /[a-zA-Z]/ ) {
    print "It's an asciibetical alpha character.\n";
}
if ( $char =~ /[[:alpha:]]/ ) {
   print "It's an alpha character.\n";
}
[download]

...just to name a few possibilities. See perlre for details.

Agreed, just let me add that you have to use locale; if you want these character classes to be locale-aware. This is like setlocale(LC_ALL, ""); in C, but it's only effective in the lexical scope of the declaration.

Update: the POSIX module has an isdigit function, but you have to apply it to a string, not a character code.

There's also CPAN. A quick search on "alphanumeric" revealed Data::Validate. More extensive digging revealed Scalar::Util::Numeric and Ctype (literally, the ctype.h tests).

my $alpha  = 'J';
my $num    = '3';

$alpha =~ /^ [[:alpha:]] $/x
and print 'correct';

$alpha !~ /^ [[:digit:]] $/x
and print 'correct';

$num  =~ /^ [[:digit:]] $/x
and print 'correct';

$num  !~ /^ [[:alpha:]] $/x
and print 'correct';
[download]