I need help with some recursion

[ragnarokPP]

Hi, I need something like this. I have file with numbers for example:
1 2 4 25 39 48
1 2 3 4 39
2 3 48
1 2 4 8 12
5 10 11 12 13
6
7 9
...
and I need to read first line and take all numbers in the 1st line (1,2,4,25,39,48) and save all the numbers in the 1,2,4,25,39 and 48 line.
So i will have (1,2,4,25,39,48,2,3,4,39,4,8,12), then i need remove duplicate items so i will have (1,2,3,4,8,12,25,39,48) and read others numbers in lines (1,2,3,4,8,12,25,39,48) and so on until i read all numbers in refers lines.
Then i need read remaining lines like (5,10,11,12,13) and (7,9) and do the same recursion with other lines. I write something like this:

while(my $line = <LIST>)
{
    chomp $line;
    push @array, $line;
}

for (my $i=0; $i<396; $i++)
{
@new_array=();
@numbers=();
if(exists $hash{$i})
{
    print OUT1 "$i\n";
}
else 
{
     @new_array=split('\s', $array[$i]);
     foreach my $el(@new_array)
     {
          chomp $el;
          push @numbers, split('\s+', $array[$el-1]);
     }
@numbers= uniq(@numbers);
@numbers= sort @numbers;
$new[$i]=@numbers;
foreach my $arg(@numbers)
{
     print OUT "$arg\n";
     $hash{$arg}=$arg;
}
}
}

sub uniq {
    return keys %{{ map { $_ => 1 } @_ }};
}

but i dont know how to set some recursion to do it right. Any idea? Thank you. Daniel

[tobyink]

To do recursion you really need to do your work in a function. That way, the function can call itself - recurse.

Right now the only function you have is a small helper, uniq and you're not doing the bulk of your work in a function.

Anyway, here's how I'd do it...

use strict;
use warnings;
use Data::Dumper;

# Read the data from the file...
#
my @lines = map { chomp; [split] } <DATA>;

# Arrays are indexed from 0, but we want to count lines
# from 1, so let's unshift a dummy 0 entry onto the array.
#
unshift @lines, undef;

# This is a more usual way of implementing "uniq". It
# preserves the original input order.
#
sub uniq
{
    my %seen;
    grep { not $seen{$_}++ } @_;
}

# This is the function that does most of the work. It
# calculates what graph theorists would call a "closure"
# over the graph of line references.
# 
sub line_closure
{
    # We're given an input list, which is a set of
    # lines already in the closure.
    #
    my @input  = @_;

    # Extend the list by looping through @input to
    # generate the output.
    #
    my @output = uniq map {
        # if the line referred to exists
        $lines[$_]
            # then add the numbers from that
            # line
            ? ($_, @{$lines[$_]})
            # otherwise, just keep the existing
            # number
            : ($_)
    } @input;
    
    # If the two lists match (and because we're only
    # ever adding to the list, checking the list
    # lengths is sufficient), then we've found the
    # closure, so return it.
    #
    if (@output == @input)
    {
        return @output;
    }
    
    # Otherwise, calculate the closure on the expanded
    # list.
    #
    return line_closure(@output);
}

# Get the closure starting at line 1.
#
my @results = line_closure(1);

print Dumper \@results;

__DATA__
1 2 4 
2 3 4
3 7
4 6
5 10 11 12 13
6
7 1

[ragnarokPP]

It is perfect, simple. But if i applied it for all the lines the result will be:

1,2,4,3,6,7
2,3,4,7,6,1
3,7,1,2,4,6
4,6
5,10,11,12,13
6
7,1,2,4,3,6

but i am looking for something like this:

1,2,4,3,6,7
5,10,11,12,13

[tobyink]

So don't apply it to all the lines. This seems to produce something close to your desired output:

my %seen;
for (my $i = 1; $i < @lines; $i++)
{
    next if $seen{$i};
    my @results = line_closure($i);
    print Dumper \@results;
    $seen{$_}++ for @results;
}

[Athanasius]

Hello ragnarokPP, and welcome to the Monastery!

tobyink has shown you how to solve the problem with recursion, as you requested. Here is a non-recursive solution using the module Set::Scalar:

#! perl
use Modern::Perl;
use Set::Scalar;

my $class_callback = sub { join(' ', sort { $a <=> $b } $_[0]->element
+s) };
Set::Scalar->as_string_callback($class_callback);

my @sets;

for (my $i = 0; <DATA>;)
{
    chomp;
    my $new_set = Set::Scalar->new(split /\s+/);
    my $merged  = 0;

    for my $j (0 .. $i - 1)
    {
        if ($new_set->intersection($sets[$j]))
        {
            $sets[$j] = $sets[$j]->union($new_set);
            $merged   = 1;
            last;
        }
    }

    $sets[$i++] = $new_set unless $merged;
}

print '(', $_, ")\n" for @sets;

__DATA__
1 2 4 
2 3 4
3 7
4 6
5 10 11 12 13
6
7 1

Output:

 1:30 >perl 424_SoPW.pl
(1 2 3 4 6 7)
(5 10 11 12 13)

 1:34 >

Remember, the Perl motto is TMTOWTDI (There’s More Than One Way To Do It)!

Update: The above code doesn’t merge fully on certain types of input. The following code fixes this:

#! perl
use Modern::Perl;
use Set::Scalar;

my $class_callback = sub { join(' ', sort { $a <=> $b } $_[0]->element
+s) };
Set::Scalar->as_string_callback($class_callback);

my @sets;

for (my $i = 0; <DATA>; ++$i)
{
    chomp;
    $sets[$i] = Set::Scalar->new(split /\s+/);
}

print "Before merging:\n";
print '(', $_, ")\n" for @sets;
print "\n";

for my $i (reverse 1 .. $#sets)
{
    for my $j (0 .. $i - 1)
    {
        if (defined $sets[$i] &&
            defined $sets[$j] &&
            $sets[$i]->intersection($sets[$j]))
        {
            $sets[$j] = $sets[$i]->union($sets[$j]);
            $sets[$i] = undef;
        }
    }
}

@sets = grep { defined } @sets;

print "After merging:\n";
print '(', $_, ")\n" for @sets;

__DATA__
1 2 4
7 13
3 5 6
7 8
10 11 12
1 5

Output:

11:24 >perl 424_SoPW.pl
Before merging:
(1 2 4)
(7 13)
(3 5 6)
(7 8)
(10 11 12)
(1 5)

After merging:
(1 2 3 4 5 6)
(7 8 13)
(10 11 12)

12:17 >

Hope that helps,

[ragnarokPP]

Thank you Athanasius. I tested it and works great. But previous code was a little bit faster in computation :). I learned something new today. Thank you guys.