in reply to Let's unite sets

I'm not sure what's going on with your "equals" concept there. Why is that approach better than 
sub union
{
    my %h;
    @h{@$_} = @$_ for @_;
    values %h
}

[download]
???

Replies are listed 'Best First'.
Re^2: Let's unite sets
by songahji (Friar) on Mar 28, 2005 at 21:06 UTC
    Would your union works for objects? 
    my @unique = union{ $_[0]->isSameNode } 
                [ qw( list of xml nodes ) ], 
                [ qw( other list of xml nodes ) ] 
                #unique node list

    [download]

    Edited by castaway - added code tags

[reply]
[d/l]

      Certainly. Notice that the hash keys are "stringified objects", which include a memory locations, like this: 

      % perl -le 'print bless +{}, "Foo"'        
Foo=HASH(0x814cbb8)

      [download]
      So objects are considered equal only if their "stringifications" match. You still recover the original object from the corresponding hash value.

      the lowliest monk

[reply]
[d/l]
      It would, and does. I encourage you to try it for yourself. I also encourage you to look at existing modules, such as Set::Object, which generally do a lot more than just union.
[reply]
        Ok it works for objects. Then how about reference object?
[reply]
Re^2: Let's unite sets
by songahji (Friar) on Mar 31, 2005 at 20:58 UTC
    are you sure it's @h{@_} not $h{@_}
[reply]
      Oh, yes, quite sure. It's called a "hash slice". You can look it up. :-)

      In a nutshell: Let's say you have an array @a; then doing 

      @h{@a} = @a;

      [download]
      is the same as doing 
      for ( @a ) {
  $h{$_} = $_;
}

      [download]
[reply]
[d/l]
[select]
        kewl!
[reply]

In Section Cool Uses for Perl