You are not counting characters* but digits \d which might be followed by 0 or more non-numbers \D* .

Not sure why?

I'd suggest better code if I new what your goal is.

Edit

The following example shows how to match (and consequently count) digits or any other character class.

No need for weird negated character classes.

  DB<6> x $_='12a3b';/\d/g
0  1
1  2
2  3
  DB<7> x $_='12a3b';/[ab]/g
0  'a'
1  'b'
[download]

*) Character_(computing) include letters, numerical digits, common punctuation marks (such as "." or "-"), and whitespace.

Update

In case you are only counting single characters and not combinations consider using tr instead, like described in Re: Some odd ambiguity in this regex