use strict;
use warnings;

use Data::Dumper;

my $rawData = <<END_OF_DATA;
species|Caragana arborescens | genus|Caragana 
| subfamily|Papilionoideae | family|Fabaceae 
| order|Fabales | no rank|eurosids I | subclass|Rosidae 
| no rank|core eudicots | no rank|eudicotyledons 
| no rank|Magnoliophyta | no rank|Spermatophyta 
| no rank|Euphyllophyta | no rank|Tracheophyta 
| phylum|Embryophyta | no rank|Charophyta/Embryophyta group 
| no rank|Streptophyta | kingdom|Viridiplantae 
| superkingdom|Eukaryota | no rank|cellular organisms 
| no rank|root || species|syncytium endosymbiont of Diaphorina citri 
| no rank|unclassified beta proteobacteria (miscellaneous) 
| no rank|unclassified beta proteobacteria | class|beta subdivision 
| phylum|Proteobacteria | superkingdom|Bacteria 
| no rank|cellular organisms | no rank|root 
|| subspecies|Trypanosoma brucei brucei | species|Trypanosoma brucei 
| subgenus|Trypanozoon | genus|Trypanosoma | family|Trypanosomatidae 
| order|Kinetoplastida | no rank|Euglenozoa | superkingdom|Eukaryota 
| no rank|cellular organisms | no rank|root 
|| species|unculturable Mariana archaeon no. 1 
| no rank|environmental samples | no rank|unclassified Crenarchaeota 
| kingdom|Crenarchaeota | superkingdom|Archaea 
| no rank|cellular organisms | no rank|root 
|| species|Suillus aeruginascens | genus|Suillus | family|Boletaceae 
| order|Boletales | subclass|Hymenomycetidae | class|Hymenomycetes 
| phylum|Basid};
END_OF_DATA
$rawData =~ s{\n}{}g;

my @rawRecords = split m{\s*\|\|\s*}, $rawData;

print scalar @rawRecords, qq{ records found\n};

my %parsedRecords;

foreach my $rawRecord ( @rawRecords )
{
    my ( $species ) = $rawRecord =~ m{(?<!\w)species\|(.+?)(?=\s+\|)};
    print qq{Species: $species\n};
    my @dataPairs = split m{\s+\|\s+}, $rawRecord;
    foreach my $dataPair ( @dataPairs )
    {
        my ( $key, $value ) = split m{\|}, $dataPair;
        unless ( exists $parsedRecords{ $species }->{ $key } )
        {
            $parsedRecords{ $species }->{ $key } = $value;
        }
        elsif ( ref $parsedRecords{ $species }->{ $key } eq q{ARRAY} )
        {
            push @{ $parsedRecords{ $species }->{ $key } }, $value;
        }
        else
        {
            $parsedRecords{ $species }->{ $key } =
               [ $parsedRecords{ $species }->{ $key }, $value ];
        }
        
    }
}

my $dd =
   Data::Dumper->new( [ \ %parsedRecords ], [ q{*parsedRecords} ] );
$dd->Indent( 1 );
print $dd->Dumpxs;