You can read each block of lines (regardless of how many lines there are) by setting the input record separator to "END DATA\n" so that each read operation reads multiple lines. Then each block can be split on newline.

use strict;
use warnings;

use Data::Dumper;

open my $inFH, q{<}, \ <<END_OF_FILE or die qq{open: $!\n};
START DATA
1 2 3 4 5 6 7
a bb ccc ddd
9 8 7 6 5
END DATA
START DATA
4.9 5.3 9.1
biff baff boff
END DATA
END_OF_FILE

my @block = ();
local $/ = qq{END DATA\n};

while ( <$inFH> )
{
    my @lines = split m{\n};
    push @block, [ @lines[ 1 .. $#lines - 1 ] ];
}

close $inFH or die qq{close: $!\n};

print Data::Dumper->Dumpxs( [ \ @block ], [ q{*block} ] );
[download]

The output is:-

@block = (
           [
             '1 2 3 4 5 6 7',
             'a bb ccc ddd',
             '9 8 7 6 5'
           ],
           [
             '4.9 5.3 9.1',
             'biff baff boff'
           ]
         );
[download]

I hope this is of use.

Cheers,

JohnGG