The first task here is to work backwards from your Dumper output to something like you have in the actual source code.
In the future, please show us the source code.

I created 3 arrays of references to hash.
When Dumper, dumps those arrays it uses $VARx=reference to array instead of actual names in your source code.
The below shows one possibility.
"Merge" is not right description here, "combine" seems more descriptive.

#!/usr/bin/perl

use strict;
use warnings;
use Data::Dumper;

my @x = (
         {
           'date' => '2001-06-04',
           'number' => '12345',
           'amount' => '100.00',
           'status' => 'paid',
           'type' => 'new'
         },
         {
           'date' => '2000-001-02',
           'number' => 'xc234',
           'amount' => '30.88',
           'status' => 'new',
           'type' => 'cost'
         }
       );

my @y = (
     {
       'ppay' => 'Smith Doe'
     }
   );

my @z = (
       {
         'deb1' => '0',
         'cred' => '0',
         'addr' => '100 - Main Street',
         'total' => '250.00 usd',
       },
       {
         'deb1' => '0',
         'cred' => '50.14',
         'addr' => '1 - Central',
         'total' => '51.00',
       }
     );
 
print "This replicates your dumper output:\n";
print Dumper \@x,\@y,\@z;

#This is what you need in your source code:
my @combined = (@x,@y,@z);

print "\n####### Desired Dumper Output ####\n";
print Dumper \@combined;

__END__
This replicates your dumper output:
$VAR1 = [
          {
            'type' => 'new',
            'status' => 'paid',
            'amount' => '100.00',
            'number' => '12345',
            'date' => '2001-06-04'
          },
          {
            'number' => 'xc234',
            'date' => '2000-001-02',
            'amount' => '30.88',
            'status' => 'new',
            'type' => 'cost'
          }
        ];
$VAR2 = [
          {
            'ppay' => 'Smith Doe'
          }
        ];
$VAR3 = [
          {
            'addr' => '100 - Main Street',
            'total' => '250.00 usd',
            'cred' => '0',
            'deb1' => '0'
          },
          {
            'addr' => '1 - Central',
            'total' => '51.00',
            'cred' => '50.14',
            'deb1' => '0'
          }
        ];

####### Desired Dumper Output ####
$VAR1 = [
          {
            'type' => 'new',
            'status' => 'paid',
            'amount' => '100.00',
            'number' => '12345',
            'date' => '2001-06-04'
          },
          {
            'number' => 'xc234',
            'date' => '2000-001-02',
            'amount' => '30.88',
            'status' => 'new',
            'type' => 'cost'
          },
          {
            'ppay' => 'Smith Doe'
          },
          {
            'addr' => '100 - Main Street',
            'total' => '250.00 usd',
            'cred' => '0',
            'deb1' => '0'
          },
          {
            'addr' => '1 - Central',
            'total' => '51.00',
            'cred' => '50.14',
            'deb1' => '0'
          }
        ];
[download]

Update: As a side note: The order that hash keys appear are random. And in more recent Perl's that is guaranteed to be so to prevent certain types of security problems. There is a "fudge factor" that gets added into the hashing function that changes per run of even the same Perl program.