Using Perl to count non-special characters in a string

Question

I am trying to have Perl count only the non-special characters (as in what one can find on a US keyboard: numbers letters, %$#() , etc.)

my $stg = "é§I¥3§";

For the string above, I have tried several examples from threads such as these: Use Perl to check if a string has only English characters

my $char_count = () = $stg =~ /[[:alpha:]]/;
my $char_count = () = $stg =~ /[a-zA-Z0-9_]/;
my $char_count = () = $stg =~ /[^a-zA-z0-9_]*$/;

All of them give me 1 when the answer I am looking for is: 2 (only 'I' and '3' counted). Can anyone tell me what the correct solution is?

Answer 1

Transliteration is another method for counting. With the /c flag, it uses the complement of the characters you specify:

my $non_ascii_count = $stg =~ tr/a-zA-Z0-9_//c;

Answer 2

You always get 1 because your regular expression stops after the 1st match. You will get 2 if you add the //g modifier ("Match globally") to your last 2 regexes:

my $char_count = () = $stg =~ /[a-zA-Z0-9_]/g;

my $char_count = () = $stg =~ /[^a-zA-z0-9_]*$/g;

The :alpha: regex only matches I , not 3 . It will return 1 even with //g :

alpha  Any alphabetical character (e.g., [A-Za-z]).