Regex a file with names and numbers/hexes alternated

Have a logfile my_file.txt which I know has the following structure:

milk marta 1231 joe 1223 ralph 7C3D
cookie
jam
milk marta 163 joe 23 ralph 7FFF

And what I want to do is extract with the help of regex in perl all the values of lines that start with the delimiter milk (which I know beforehand). Then I know I will have a number of value names (in the toy example marta, joe, ralph), each followed by a decimal or hexadecimal value.

• Is there a way to read the values without knowing beforehand if they are decimal or hex?

• What is the best way to store this values? I thought of doing a hash with the keys being marta, joe, ralph, but I might not know them beforehand, meaning I will know all of them once I have read the first line, not before.

And here comes my attempt so far

$my_file = "my_file.txt"
open my $fh, $my_file or die "Nope!"

while (my $line = <$fh>)
{
    if ($line =~ m/milk/)
    {
        # here iterate with a for or while loop over the $line
        # and read word and value
    }
}

Aside from the issue of determining whether a digit string is decimal or hex, something like this will extract the text that you need

use strict;
use warnings 'all';

use Data::Dump 'dd';

my @data;

while ( <DATA> ) {
    my ($col1, %item) = split;
    push @data, \%item if $col1 eq 'milk';
}

dd \@data;

__DATA__
milk marta 1231 joe 1223 ralph 7C3D
cookie
jam
milk marta 163 joe 23 ralph 7FFF

output

[
  { joe => 1223, marta => 1231, ralph => "7C3D" },
  { joe => 23, marta => 163, ralph => "7FFF" },
]