in reply to recursive difficulty

A couple of things. First and foremost you once you have read everything out of <DATA> you would not be able to call the routine again to reload the data. (a simple example should show what i mean):
use strict;
use warnings;
get_data();
get_data();

sub get_data    
{
   while ( <DATA> ) {
     print $_;
   }
}   ##get_data


__DATA__
A
B
C
D

[download]
Second, instead of popping things off of the array (in the hash) and thus changing the data in the structure, I would use a construct like the following,so that the data structure stays intact and thus there would be no reason to rebuild it. What I mean is that, instead of this:
while(1){
         last unless my $immed = pop @{$immed_child{$root}};
        get_descends($immed);
        $all_descends{$immed}++;
    }

[download]
I would do something like this instead:
    foreach my $immed ( @{$immed_child{$root}} )  {
        get_descends($immed);
        $all_descends{$immed}++;
    }

[download]
Anyhow, I hope this helps.

-enlil

Replies are listed 'Best First'.
Re: 2: recursive difficulty
by jjohhn (Scribe) on Mar 09, 2003 at 23:57 UTC
    Thank you. I did as you suggested, and it seems like my split has now somehow broke: I get the error "Use of unitialized value in concatenation (.) or string at tree.pl line 36, <DATA> line 1 (for all the <DATA> lines).

    I copied and pasted the code from perlmonks onto a different computer; that may be related.

    I thought perhaps my vim expandtabs option on this machine was involved, but I retyped the DATA using ^V. This code

    # key: parent 
# val: list of immediate children
sub load_hashes{
    my ($child, $par);
    while(<DATA>){
        chomp;
        ($child, $par) = split /"\t"/;
#push @{$immed_child{$par}}, $child;
        print "ch: $child, par: $par\n";
    }
}

    [download]
    prints the entire table row as the value of $child, and no value for $par. For example:

    ch: jumprope exercise_device, par:

    this explains the unitialized value in the string. Is there an obvious mistake in my split that has somehow occurred in the copying of the code from perlmonks onto a new machine?

[reply]
[d/l]
      That's one of the problem I had at the beginning. By copy and paste, all those \t's are gone. I just manually added those \t's back to your __DATA__. Those \t's are what you used in your split.
[reply]
        That seems not to be working for me, despite using ^V. I think it is my vim expandtab setting. I tried pipes as delimiters, but split /|/ seemed to split between characters for some reason:

        ch: v, par: e
        in my little debug printout.

        I am a litttle baffled.

[reply]

In Section Seekers of Perl Wisdom