Actually, it's not clear from your code whether you need the array to be an object, too. Do you have methods that use the array as a whole? It's perfectly fine to have an array of objects.

Also, this line would never work:

my (@class, $count) = @_;
[download]

You should call it with $itemref, this is the variable holds the object. Even if this was the case, you would get error because in your method definition, you use an array reference but dereference it as HASH.

The $_[0] is the first element of @_, which holds parameters passed in to methods/subroutines. In a method, $_[0] is supposed to be the object itself, hence the use of $self as a convetional variable name.

I assume that you (hopefully) have a sort of this structure:

# simplified version
my @array = ({id=>1, name=>'Linux'}, {id=>2, name=>'Perl'});
my $self = [@array];
bless $self, $class
[download]

Remember, the first argument to method is the object itself. Your object is a reference to an ARRAY so you can't dereference it as HASH. Instead, you want something like this.

sub get_desc {
  my($self, $idx) = @_;
  return $self->[0]{_description};
}
[download]

And you call it like this,

my $item = Item::Item->new('some record id');
print $item->get_desc(0);
[download]

You should stay with your object after creating it, don't dereference to use it to call an object method.

my $itemref = Item::Item->new( "10056" );
[download]

@item = @$itemref;
print $item[0]->{_description};
[download]

gave you the right thing because you treated it as data structure. But remember, your object is $itemref. So stick with it whenever you want to use it as object.

When I first learned about OOP (object oriented Perl :-), I found perlobj and perltoot as good introductions and enjoyful to read.

Thanks to all, for all the help (I have been learning perl for a couple of months - all prior work procedural, and like it). I did not realise I would need to bless slices of an array to. I don't want to move the sql statement outside of the package so it calls the bless function six times (correct me if I am wrong) but have it all contained (except the where part of the query) in the package. So I followed your code, changing

return bless \@objref, $class;
to
$self = [@objref];
return bless $self, $class;

and changed the sub to 
sub get_desc()
 {
  my ($self, $idx) = @_ ;
  return $self->[$idx]{_description} ;
 }
[download]

calling code from

my $item = Item::Item->new( "10056" );
print $item->get_desc(0), "\n";
[download]

it works!. However when I try the next entry (the sql statement returns 6 rows)

my $item = Item::Item->new( "10056" );
print $item->get_desc(1), "\n";
[download]

I get: Use of uninitialized value in print at testitem.pl line 17. In my while statement I assign each row to the $objref[$i++] slice. The $self = [@objref] should have the whole array reference (handle ? - not sure) and so why does it not work? More reading to do...

.

Sorry my mistake - this query only returns 1 row. I tried it with success on a multi-row query. Thank you.

Perhaps something like this:

#!/usr/bin/perl

use strict;
use warnings;

{
  package Item;
  
  sub new {
    my ($class) = @_;
    my $self = [
      {fld1 => 1, fld2 => 2},
      {fld1 => 3, fld2 => 4},
    ];
    bless $self, $class;
  }
  
  sub get_fld {
    my ($self, $rec, $fld) = @_;
    return $self->[$rec]{$fld};
  }
  
}

my $item = Item->new;
my $fld = $item->get_fld(1, 'fld1');
print "$fld\n";
[download]

If you have not blessed your reference, your program will not know where it is going to find the get_desc() method.

Take a look at this tutorial on how to create perl objects

(Oh, and one more hint: ignore whatever you hear about "Inside-Out Objects" for a little while longer. There's still room for debate about what the right way is to implement those, and most OOP perl code you see will be hash based).