Recursively printing data structures in Perl

Question

I am currently learning Perl. I have Perl hash that contains references to hashes and arrays. The hashes and arrays may in turn contain references to other hashes/arrays.

I wrote a subroutine to parse the hash recursively and print them with proper indentation. Though the routine works as expected, my instructor was not convinced about the readability and elegance of the below code.

I would really appreciate to get the views of Perl experts here on possible optimization of the below code.

Here is my complete code snippet..

# Array of Arrays
$ref_to_AoA = [
     [ "fred", "barney" ],
     [ "george", "jane", "elroy" ],
     [ "homer", "marge", "bart" ],
];


#Array of Hashes
$ref_to_AoH = [ 
{
   husband  => "barney",
   wife     => "betty",
   son      => "bamm bamm",
},
{
   husband => "george",
   wife    => "jane",
   son     => "elroy",
},
];

# Hash of Hashes
$ref_to_HoH = { 
    flintstones => {
        husband   => "fred",
        pal       => "barney",
    },
    jetsons => {
       husband   => "george",
       wife      => "jane",
       "his boy" => "elroy",  # Key quotes needed.
    },
    simpsons => {
       husband   => "homer",
       wife      => "marge",
       kid       => "bart",
    },
};

# Hash which contains references to arrays and hashes
$finalHash = {
   'arrayofArrays' => $ref_to_AoA,
   'arrayofHash' => $ref_to_AoH,
   'hashofHash' => $ref_to_HoH,
};

$string = str($finalHash); 
print "$string\n";

#------------------------------------------------------------------
sub str {
    my $hash = shift;
    my ($space, $newline, $delimiter) = @_;
    $space = "" unless (defined $space);
    $newline = "\n\n\n" unless (defined $newline);
    $delimiter = "\n--------------------------------------------" unless (defined $delimiter);
    my $str = "";

    for (sort keys %{$hash}) {
        my $value = $hash->{$_};
        $str .= "$newline$space$_ == $value$delimiter";
        $str .= recurseErrors($value,$space);
    }
    $str;
}

#------------------------------------------------------------------
sub recurseErrors {
    my $str;
    my ($value,$space) = @_;
    my $ref = ref $value;

    if ($ref eq 'ARRAY') {
        my $i = 0;
        my $isEmpty = 1;
        my @array = @$value;
        $space .= "\t";
        for my $a (@array) {
            if (defined $a) {
                $isEmpty = 0;
                $str .= "\n$space$_\[$i\] :";
                $str .= recurseErrors($a,$space);
            }
            $i++;
        }
        $str .= "= { }" if ($isEmpty);

    } elsif ($ref eq 'HASH') {
        $space .= "\t";
        for my $k (sort keys %$value) {
            if ( ( ref($value->{$k}) eq 'HASH') || (ref $value->{$k} eq 'ARRAY') ) {
                my $val = $value->{$k};
                $str .= "\n\n$space$k == ";
                $str .= "$val";
            }
            else {
                $str .= "\n$space$k == ";
            }
            $str .= recurseErrors($value->{$k},$space);
      }

      # we have reached a scalar (leaf)
    } elsif ($ref eq '') {
        $str .= "$value";
    }
$str
}
#------------------------------------------------------------------

Output:

arrayofArrays == ARRAY(0x9d9baf8)
--------------------------------------------
    arrayofArrays[0] :
        arrayofArrays[0] :fred
        arrayofArrays[1] :barney
    arrayofArrays[1] :
        arrayofArrays[0] :george
        arrayofArrays[1] :jane
        arrayofArrays[2] :elroy
    arrayofArrays[2] :
        arrayofArrays[0] :homer
        arrayofArrays[1] :marge
        arrayofArrays[2] :bart


arrayofHash == ARRAY(0x9d9bba8)
--------------------------------------------
    arrayofHash[0] :
        husband == barney
        son == bamm bamm
        wife == betty
    arrayofHash[1] :
        husband == george
        son == elroy
        wife == jane


hashofHash == HASH(0x9da45f8)
--------------------------------------------

    flintstones == HASH(0x9d9bb48)
        husband == fred
        pal == barney

    jetsons == HASH(0x9d9bbf8)
        his boy == elroy
        husband == george
        wife == jane

    simpsons == HASH(0x9d9bc48)
        husband == homer
        kid == bart
        wife == marge

Answer 1

I've struggled with this same problem before, and found my way here. I almost used a solution posted here, but found a more suitable one (for me anyway). Read about Depth First Recursion here.

The sub in the above article works perfectly with a reference containing other Hashes, Arrays, or Scalars. It did not print Hash key names, though, so I slightly modified it:

#!/usr/bin/perl
#
# See:
#
# http://perldesignpatterns.com/?DepthFirstRecursion
#
use strict;
use warnings;

my %hash = (
  'a' => {
    'one' => 1111,
    'two' => 222,
  },
  'b' => [ 'foo', 'bar' ],
  'c' => 'test',
  'd' => {
    'states' => {
      'virginia' => 'richmond',
      'texas' => 'austin',
    },
    'planets' => [ 'venus','earth','mars' ],
    'constellations' => ['orion','ursa major' ],
    'galaxies' => {
      'milky way' => 'barred spiral',
      'm87' => 'elliptical',
    },
  },
);

&expand_references2(\%hash);

sub expand_references2 {
  my $indenting = -1;
  my $inner; $inner = sub {
    my $ref = $_[0];
    my $key = $_[1];
    $indenting++;
    if(ref $ref eq 'ARRAY'){
      print '  ' x $indenting,'ARRAY:';
      printf("%s\n",($key) ? $key : '');
      $inner->($_) for @{$ref};
    }elsif(ref $ref eq 'HASH'){
      print '  ' x $indenting,'HASH:';
      printf("%s\n",($key) ? $key : '');
      for my $k(sort keys %{$ref}){
        $inner->($ref->{$k},$k);
      }
    }else{
      if($key){
        print '  ' x $indenting,$key,' => ',$ref,"\n";
      }else{
        print '  ' x $indenting,$ref,"\n";
      }
    }
    $indenting--;
  };
  $inner->($_) for @_;
}

Answer 2

    #use strict ; 
    use warnings ; 
    # use module
    use XML::Simple;
    use Data::Dumper;

    #debug print "START SCRIPT " ; 

    my $fileToParse = 'C:/Temp/CDIP/scripts/perl/nps_all_workflows.xml' ; 

    # create object
    my $objXml= new XML::Simple;

    # read XML file
    my $data = $objXml->XMLin("$fileToParse");

    # #debug print "\n FirstLevel is " . $objXml->{'POWERMART'} ; 
    my $level = 1 ; 

    #
    printHashKeyValues ($data ) ;  


    sub printHashKeyValues  
    {
        $level ++ ; 
        my $refHash = shift ; 
        my $parentKey = shift ; 
        my $parentValue = shift ; 


        while( my ($key, $value) = each %$refHash) 
        {


                        if ( defined ( $key ) )
                        {
                                if ( ref ($refHash->{"$key"}) eq 'HASH'  ) 
                                 {
                                 my $newRefHash = $refHash->{"$key"} ; 
                                 #debug print " \n The key is a hash " ; 
                                 printHashKeyValues ($newRefHash , $key , $value) ;
                                 }


                                    if ( ref ($refHash->{"$key"}) eq 'ARRAY'  ) 
                                     {
                                            #debug print " \n the key is an ARRAY " ; 
                                            printArrayValues ( $refHash->{"$key"} ) ; 
                                     }


                        } #eof  if ( defined ( $key ))

                        if ( defined ( $value) )
                        {

                                if (  ref ($refHash->{"$value"}) eq 'HASH'  ) 
                                 {
                                 my $newRefHash = $refHash->{"$value"} ; 
                                #debug print " \n The value is a hash " ; 
                                 printHashKeyValues ($newRefHash , $key , $value) ;
                                 }


                                if ( ref ($refHash->{"$value"}) eq 'ARRAY'  ) 
                                 {
                                        #debug print " \n the value is an ARRAY " ; 
                                        printArrayValues ( $refHash->{"$value"} ) ; 
                                 }

                         } #eof if defined ( $value ) 

                                #debug print "\n key: $key, value: $value.\n";


        } #eof while
    } #eof sub 



    sub printArrayValues
    {
        my $arrRef = shift ; 
        my @array = @$arrRef; 
        my $parrentArrayElement = shift ; 

            #debug print "printArrayValues CALLED " ;

        foreach my $arrayElement ( @array ) 
        {
                        if (defined ( $arrayElement ) )
                        {
                                if   ( ref ($arrayElement) eq 'HASH'  ) 
                             {
                                 #debug print " \n The  \$arrayElement is a hash FROM THE ARRAY " ; 
                                 printHashKeyValues ($arrayElement ) ;  
                             } #eof if 

                                if   ( ref ($arrayElement) eq 'ARRAY'  ) 
                             {
                                 #debug print " \n The \$arrayElement is a ARRAY  FROM THE ARRAY " ; 
                                 printArrayValues ($arrayElement ) ;  
                             } #eof if 

                                #debug print "\n \$arrayElement is $arrayElement " ; 
                        } #eof if ( defined ( $arrayElement ) ) 



        } #eof foreach 

    } #eof sub 




    # #debug print output
    ##debug print Dumper($data);




    1 ; 

Answer 3

Here is one simple example why your code is not easily readable:

$delimiter = "\n--------------------------------------------" unless (defined $delimiter);

You could use the defined or operator:

$delimiter //= "\n" . '-' x 44;

If you are worried about earlier Perls:

defined $delimeter or $delimeter = "\n" . '-' x 44;

Conditionals going off the right margin are enough of a turn-off for me not to read the rest of the code.

Answer 4

1. Always use use strict;
2. To be a good boy, use use warnings as well.
3. The names you use for subroutines should make it obvious what the subroutine does. "recurseErrors" kind of violates that principle. Yes, it does recurse. But what errors?
4. On the first line of each subroutine you should declare and initialize any parameters. recurseErrors first declares $str and then declares its parameters.
5. Don't mix shift and = @_ like you do in str()
6. You might consider breaking up what is now called recurseErrors into specialized routines for handling arrays and hashes.
7. There's no need to quote variables like you do on lines 99 and 109.

Apart from that I think your instructor had a bad day that day.

Answer 5

You could have separated out the code blocks that dealt with arrays, and hashes.

sub recurse{
  ...
  recurse_A(@_) if $ref eq 'ARRAY';
  recurse_H(@_) if $ref eq 'HASH';
  ...
}

sub recurse_A{ ... }
sub recurse_H{ ... }

I would recommend starting out your subroutines like this, unless you have a real good reason for doing otherwise.

sub example{
  my( $one, $two, $three, $optional_four ) = @_;

( If you do it like this then Komodo, at least, will be able to figure out what the arguments are to your subroutine )

There is rarely any reason to put a variable into a string containing only the variable.

"$var" eq $var;

The only time I can think I would ever do that is when I am using an object that has an overloaded "" function, and I want to get the string, without also getting the object.

package My_Class;
use overload
  '""' => 'Stringify',
;
sub new{
  my( $class, $name ) = @_;
  my $self = bless { name => $name }, $class;
  return $self;
}
sub Stringify{
  my( $self ) = @_;
  return $self->{name};
}

my $object = My_Class->new;
my $string = "$object";

Answer 6

If you are new to perl, I'd recommend running your code through perl-critic (there is also a script you can install from CPAN, normally I use it as a test so it gets run from the command line whenever I do "make test"). In addition to its output, you might want to break up your functions a bit more. recurseErrors has three cases that could be split into sub functions (or even put into a hash of ref-type to sub-function ref).

If this were a production job, I'd use Data::Dumper, but it sounds like this is homework, so your teacher might not be too pleased.