Vim wordcount for alphanumeric characters only

Question

I know that vim has a function wordcount() that returns a dictionary containing the number of bytes, lines, words etc in the current buffer. The problem I have is that the words field of the dict from wordcount() seems to include the number of WORDs (vim's term for any sequence of chars that are not a whitespace character). For example, given:

# a markdown title

in a buffer, wordcount()['words'] returns 4, because the "#" counts as a word to vim. Is there some way I can count only the alphanumerical words in a buffer? I know I can use

:%s/\w\+//eng

but redirecting the output is cumbersome and slow if I want to have a constant word count (say, in the statusline). I need this for writing assignments.

Answer 1

You can split() on non-word characters and count the number of resulting words.

:let cnt=0 | g/./let cnt+=len(split(getline('.'), '\W\+'))