in reply to Counting the frequency of words in a string

I think the hardest part would be to define what makes "a word".

Given the definition was everything that doesn't match \W, this could do it:

my ( %wordcount, $relevancy );
for ( split /\W+/, $string ) # gets a list of all words
{
    $wordcount{$_}++;
}

foreach my $word ( @list_of_words )
{
    $relevancy += $wordcount{$word};
}

[download]

I know there are smarter ways, for this can consume a lot of memory. Read on... =)

Cheers, Sören

Replies are listed 'Best First'.
Re: Counting the frequency of words in a string
by cez (Novice) on Dec 02, 2004 at 22:12 UTC
    how about:

    use Regexp::List, put the generated regex into $word_regex

    my $count = ($count = $text) =~ s/$word_regex//g;

    horrible?

[reply]
      oh, i guess that could be just

      my $count = $text =~ s/$word_regex//g;

      (since destroying $text doesn't matter in my case)

[reply]
        Or 
        my $count = () = $text =~ /$word_regex/g;

        [download]
        The parens make it list context, then you take the scalar context from that.
[reply]
[d/l]

In Section Seekers of Perl Wisdom