$string =~ /((\w|'|-)+)/g
[download]

$string =~ /([-\w']+)/g
[download]

$string =~ /(\w+[-']?\w+)/g
[download]

$string =~ /(\w+[-']?\w+|\w)/g
[download]

#!/usr/local/bin/perl

use strict;
use warnings;

$/ = undef;
my $string = <DATA>;
while ($string =~ /(\w+[-']?\w+|\w)/g) {
        print "Word: <$1>\n";
}
__DATA__
This is a sentence with words that're different from other words.  
They have apostrophes in them (') and dashes, or dash-like characters
(-).
[download]

The ultimate guide (in my opinion) on regular expressions is Jeffrey Friedl's Mastering Regular Expressions, 2nd Edition.

Arjen

Thanks, I'll run the code now. There are so many possible combination of word into a single composed word, that I need to test every case.

while ($string =~ /((\w|')+)/g) {
[download]

((\w|')+) matches one or more occurences of (\w|'). Because of the first left paren, this captures into $1. $2 will be the same as the last char of $1.

//g in a scalar context returns the next match, so the while will continue until all matches are found.

If you just want to add hypenated words, try this:

while ($string =~ /(([\w'-])+)/g) {...do something...}
[download]

Update: To get everything surrounded by whitespace, use

while ($string =~ /((\S)+)/g) {...do something...}
[download]

You might as well use a character class since you aren't being too discriminatory about the words you match (i.e., your RE will also match a "word" like foo'bar'baz)

while ($string =~ /([\w'-]+)/g) { ... }
[download]

Note that you have to put the dash character either as the first or last character in the character class so that it isn't interpretted as a range

/('?[\w-]+'?)/g
[download]

I want my string to accept composed word which have a character "-" between them such as "tear-down" or by-the-house; I don't need to be surrounded by quotes... maybe to get everything separate by blanks in a sentence?

So then why did you put the quotes in your regex?

Never mind. It is important to remember that the Perl regexps are greedy: they will match as much as possible. Therefore to match a word, you only need to specify what a word is, there is no need to define what it's delimiters might be.

That simplifies the regexp to:

/([\w-]+)/g
[download]

use YAPE::Regex::Explain;
print YAPE::Regex::Explain->new(qr/((\w|')+)/)->explain();
__DATA__
The regular expression:

(?-imsx:((\w|')+))

matches as follows:
  
NODE                     EXPLANATION
----------------------------------------------------------------------
(?-imsx:                 group, but do not capture (case-sensitive)
                         (with ^ and $ matching normally) (with . not
                         matching \n) (matching whitespace and #
                         normally):
----------------------------------------------------------------------
  (                        group and capture to \1:
----------------------------------------------------------------------
    (                        group and capture to \2 (1 or more times
                             (matching the most amount possible)):
----------------------------------------------------------------------
      \w                       word characters (a-z, A-Z, 0-9, _)
----------------------------------------------------------------------
     |                        OR
----------------------------------------------------------------------
      '                        '\''
----------------------------------------------------------------------
    )+                       end of \2 (NOTE: because you're using a
                             quantifier on this capture, only the
                             LAST repetition of the captured pattern
                             will be stored in \2)
----------------------------------------------------------------------
  )                        end of \1
----------------------------------------------------------------------
)                        end of grouping
----------------------------------------------------------------------
[download]