If you check How do I determine whether a scalar is a number/whole/integer/float? you'll see that there are several different regex to choose from depending on what types of numbers you want to allow.

uhm, well. The numbers with comma's as delimiters are just there and just can't seem to get to 5.3 as the result of the calculation. Well, we'll just go on and a hint is appreciated.
Looked at the regex and now asking myself how to implement that.

okay I'm there for the moment. Thank you all.
I finally put the regex, I was adviced to use, in the place where it works. Code below could be improved because when starting the first column with a 'space' or nothing there won't show up a 0 (zero) in the column but that also wasn't really where I was after.
Thangs again, rethinking the issue made me learn and get to a solution.

#!/usr/bin/perl
use strict;
use warnings;
use Data::Table;

my $t =
  new Data::Table( [ [ "1", "2", "3" ], [ "10", "20", "2,3" ] ], [ "A"
+, "B" ],
    1 );

# create emtpy column
my @newCol = ();
$newCol[ $t->nofRow - 1 ] = undef;
$t->addCol( \@newCol, 'adding_up' );

# to change , into . in respective columns
# not needed for a correct calculation of col3 though
$t->colsMap( sub { $_->[0] =~ s/,/\./g } );
$t->colsMap( sub { $_->[1] =~ s/,/\./g } );

# Return input if input is an integer, otherwise zero
sub to_int_or_zero {
    my $value = shift;
    if (
        defined($value)

        # the regex that above hints pointed at.
        && $value =~ /^-?\d+\.?\d*$/
      )
    {
        return $value;
    }
    else {
        return 0;
    }
}

# Put the formula in the empty column
$t->colsMap(
    sub {
        $_->[2] = to_int_or_zero( $_->[0] ) + to_int_or_zero( $_->[1] 
+);
    }
);
print $t->csv;
[download]