Name Me! MixMatch?

jackdied has asked for the wisdom of the Perl Monks concerning the following question:

Here is a module I use for solving little problems (it's in my Dr Dobbs toolkit) and I need a better name. Someday I might throw it in with my Combination module and Permutation module (both of which I also use for Dr Dobbs type problems), and then I'll need to figure out a namespace. For now, I just need to name this little guy.

package MixMatch;
use strict;
 
=comments
 
Takes an lref of lrefs like
 
[
  [a],
  [x, y, z]
  [1, 2]
]
 
and returns in order
  [a, x, 1]
  [a, x, 2]
  [a, y, 1]
  [a, y, 2]
  [a, z, 1]
  [a, z, 2]
 
=cut
 
sub new {
    my $self = bless {}, shift;
    my %args = (-array=>undef, # lref of lrefs
                @_);
 
    $self->{arr} = $args{-array};
    my $sizes = [];
    my $total = 1;
    foreach my $lref (@{$self->{arr}}) {
        push @$sizes, scalar(@$lref);
        $total *= $sizes->[-1];
    }
    $self->{rows} = scalar(@{$self->{arr}});
    $self->{sizes} = $sizes;
    $self->{count} = $total;
    return $self;
}
 
sub next {
    my $self = shift;
    my $c = $self->{count}--;
    return undef  unless $c;
    my $ret = [];
    for (my $i = 0; $i < $self->{rows}; $i++) {
        push @$ret, $self->{arr}->[$i]->[$c % $self->{sizes}->[$i]];
    }                             
    return $ret;
}

1;
[download]

I won't post my Permutation.pm here, it is 4x slower than Algorithm::Permute (which is a nice piece of XS), I just use mine because I like the order of permutations is produces. And yes, this module could be implemented in terms of Permute's, but it is so short that new()ing a few other objects seems like overkill

Suggestions?

-jackdied

Comment on Name Me! MixMatch? Download Code

Replies are listed 'Best First'.
Re: Name Me! MixMatch? by ariels (Curate) on Dec 19, 2001 at 14:20 UTC
Your function computes the cartesian product, why not name it after that? Math::CartesianProduct, maybe? While on the subject, IWBNI you featured a tied-filehandle interface, which would let me tie a filehandle to Math::CartesianProduct, then read it instead of using your `next` method.	[reply]
Re: Re: Name Me! MixMatch? by jackdied (Monk) on Dec 19, 2001 at 14:31 UTC
Very cool. If I had only know that before, I wouldn't have had to write it :p Also, can reading from a file handle return non-scalars? -jackdied	[reply]
Re: Re: Re: Name Me! MixMatch? by ariels (Curate) on Dec 19, 2001 at 14:38 UTC
Reading from a file handle returns either a scalar, if you're doing something like "`$x = <FH>`", or a list, if you're doing something like "`@x = <FH>`". These are the only 2 contexts (scalar context and list context) available in Perl (well, there's also a void context, if you're really picky). But I don't think that's what you're asking. If you meant "can reading from a file handle return non-strings", the answer is also positive. Your `READLINE` method can return any scalars it likes. A list reference is just a scalar, so it could return a list reference in scalar context and a list of list references in list context. Just like any Perl function, really.	[reply] [d/l] [select]
Re (tilly) 1: Name Me! MixMatch? by tilly (Archbishop) on Jan 02, 2002 at 04:40 UTC
An alternate approach for the same kind of problem is to use functional techniques as at Re (tilly) 1 (perl): What Happened...(perils of porting from c). Also note that if you have a large number of possible options to go through, this will be less memory intensive.	[reply]
Broken, except for trivail cases (such as the example) by jackdied (Monk) on Dec 20, 2001 at 04:28 UTC
This code is broken, and only does the right thing because of the 1/2/3 nature of the example. Use this patch (only a few lines have changed, so it should be human readable) to get it to work for all cases. RCS file: /home/perl/tools/Cartesian.pm,v retrieving revision 1.2 diff -r1.2 Cartesian.pm 30c30,31 < my $sizes = []; --- > my $div = []; > my $mod = []; 33,34c34,36 < push @$sizes, scalar(@$lref); < $total = $sizes->[-1]; --- > push @$div, $total; > push @$mod, scalar(@$lref); > $total = $mod->[-1]; 37,38c39,42 < $self->{sizes} = $sizes; < $self->{count} = $total; --- > $self->{div} = $div; > $self->{mod} = $mod; > $self->{total} = $total; > $self->{count} = 0; 44,45c48,49 < my $c = $self->{count}--; < return undef unless $c; --- > my $c = $self->{count}++; > return undef if ($c == $self->{total}); 48c52 < push @$ret, $self->{arr}->[$i]->[$c % $self->{sizes}->[$i]]; --- > push @$ret, $self->{arr}->[$i]->[($c / $self->{div}->[$i]) % + $self->{mod}->[$i]]; [download] Woops	[reply] [d/l]
Re: Name Me! MixMatch? by chip (Curate) on Dec 19, 2001 at 14:08 UTC
If you're the only one who's going to use it, why not give it a name only you would recognize? -- Chip Salzenberg, Free-Floating Agent of Chaos	[reply]
Re: Name Me! MixMatch? by blakem (Monsignor) on Jan 15, 2002 at 05:26 UTC
Perhaps you're unfamiliar with this glob trick? `perl -le 'print for glob("{a}:{x,y,z}:{1,2}")' a:x:1 a:x:2 a:y:1 a:y:2 a:z:1 a:z:2` [download] -Blake	[reply] [d/l]