Iterate over a nested hash in perl - perl

Question

I have to iterate over a nested hash in perl and carry out some operations. The structure I have is
$featureGroup = [
{
featureType => "widget",
name => "dpx-shadow-fleet",
parameterMap => { dpxContext => "shadowAtf", dpxEndPoint => "/art/dp/ppd?" },
},
{
featureType => "widget",
name => "dpx-shadow-fleet",
parameterMap => { dpxContext => "shadowBtf", dpxEndPoint => "/art/dp/btf?" },
},
{
features => [
{
featuredesc => [
{
critical => 1,
featureType => "widget",
name => "dpx-ppd",
parameterMap => { dpxContext => "atf", dpxEndPoint => "/art/dp/" },
},
{
featureType => "widget",
name => "error",
parameterMap => { errorMessageId => "error" },
},
],
featureType => "sequence",
},
{
critical => 1,
features => ["encode-landing-image", "image-encoding-error"],
featureType => "sequence",
},
],
handler => "/gp/product/features/embed-landing-image.mi",
name => "embed-landing-image",
pfMetrics => { "" => undef, "start" => sub { "DUMMY" }, "stop" => sub { "DUMMY" } },
type => "custom-grid",
},
];
I want to iterate over the featuredesc subarray and get the value name. I am trying out this.
for(my $i = 0; $i < #$featureGroup; $i++){
if(defined $featureGroup->[$i]->{'features'}){
for(my $j = 0; $j < #$featureGroup->[$i]->{'features'} ; $j++){
print "$featureGroup->[$i]->{'features'}->{'featuredesc}->{name}";
}
}
}
But this is not working. I am not understanding where am I going wrong. Any pointers in the right direction would be useful.

Answer 1

You have a very complex data object there and you have already encountered problems dealing with it. While I could help you address your direct problem, I think you would benefit more from learning how to reduce the complexity.
Perl supports Object Oriented programming. This allows you to take data structures and attach subroutines to them that operate on them. You can read about Perl OO here. I will show you quickly how you can turn the $featureGroup list into a list of objects, and how to access the features that a single object contains. You should apply this technique to every hash in your datastructure (you can tone it back if you are sure that certain inner hashes should not be objects, but it is probably better to start by overdoing it and then scale back rather than the other way around).
This is one of the feature group hashes:
{
'featureType' => 'widget',
'name' => 'dpx-shadow-fleet',
'parameterMap' => {
'dpxContext' => 'shadowAtf',
'dpxEndPoint' => '/art/dp/ppd?'
}
}
In this one you have a featureType, name, and parameterMap. These fields do not appear in every object in your list (in fact the last hash looks quite different to the first two). I will show you how to create an object which requires those three parameters:
package Feature;
use Moose; # You may have to install this
has 'featureType' => (
'is' => 'rw',
'isa' => 'Str'
);
has 'name' => (
'is' => 'rw',
'isa' => 'Str'
);
has 'parameterMap' => (
'is' => 'rw',
'isa' => 'HashRef'
# You could make this accept another object type
# if you convert this inner hash
);
You can then construct your object like so:
my $f = new Feature(
'featureType' => 'widget',
'name' => 'dpx-shadow-fleet',
'parameterMap' => {
'dpxContext' => 'shadowAtf',
'dpxEndPoint' => '/art/dp/ppd?'
}
);
You are then able to access those fields by using the named accessors:
print $f->name; # dpx-shadow-fleet
At the moment this just seems like a longer way to use a hash, right? Well the real benefit comes from being able to define arbitrary subroutines on the class which hide complexity from the caller. So you want to operate on the features array in your original question. Lets define that as a field:
has features => (
is => 'rw',
isa => 'ArrayRef[HashRef]'
# This is an array containing hashes
# You _really_ want to turn the inner hashes into an object here!
);
Then we can operate on them in another subroutine. Lets define one that returns every feature that is a sequence (has a featureType of sequence):
sub get_sequences {
my ($self) = #_;
return grep { $_->{featureType} eq 'sequence' } #{ $self->features };
}
Now when you use an object of this type to get the sequence features all you need to do is:
$f->get_sequences();
If you apply this to all levels of your hash you will find that your code becomes easier to manage. Good luck!

Answer 2

Try this:
for(my $i = 0; $i < #$featureGroup; $i++){
if(defined $featureGroup->[$i]->{'features'}){
for(my $j = 0; $j<scalar #{$featureGroup->[$i]->{'features'}} ; $j++){
for(my $k=0;$k<scalar #{$featureGroup->[$i]->{'features'}->[$j]->{'featuredesc'}};$k++) {
if (defined $featureGroup->[$i]->{'features'}->[$j]->{'featuredesc'}->[$k]->{'name'}) {
print $featureGroup->[$i]->{'features'}->[$j]->{'featuredesc'}->[$k]->{'name'}."\n";
}
}
last if !defined $featureGroup->[$i+1]->{'features'};
}
}
}

Answer 3

Instead of iterated by index, I'd advise that you iterate by element.
This enables one to easily filter each step using grep or next
for my $group (grep {$_->{features}} #$featureGroup) {
for my $feature (grep {$_->{featuredesc}} #{$group->{features}}) {
for my $desc (#{$feature->{featuredesc}}) {
print "$desc->{name}\n"
}
}
}
Outputs:
dpx-ppd
error