If the patterns are the same, then you actually have a simpler case (though the more complicated one shouldn't break, and is more generalized):

my @array;
while (<DATA>) { last if /^d/ } #discard until the first match
while (<DATA>) {
   if ( @array and /^d/ ) {
      process_array( @array );
      undef @array;
      next;
   }
   push @array, $_;
}
[download]

Of course, depending on the size of your data set, you could slurp and split:

my $data  = join('',<DATA>);
my @chunk = split(/^\d/m, $data);

foreach (@chunk) { process_array( split(/\n/s, $_) ) }
[download]

But this latter assumes your data starts and ends OK (won't skip headers or the like).