The solution by poj in 11100896 is cleaner as it handles the extra space at read time. However, you might only want to remove those following a comma, for which you can pass a regex to split.

push @multi_array, [split(/,\s*/, $_)] for <DATA>;
[download]

I also gave an incorrect example for the map in my previous post, as the code did not operate on array refs. The map can be replaced by code like this:

foreach my $row (@multi_array) {
    s/^\s+// foreach (@$row);
}
[download]

And it's a stylistic thing, but I would avoid using so many postfix loops, and only use the topic variable where it is clearer to do so (you might have good reasons to think differently, though). For example, this code snippet is getting difficult to read at the first pass (for me at least):

# modify columns
($_->[0] *= 1.0, $_->[1] /= 1.0) for @multi_array;
[download]

It is clearer as this (similar to what poj posted):

for my $row (@multi_array) {
    $row->[0] *= 1.0;
    $row->[1] /= 1.0;
}
[download]

And the nested foreach loops above could be:

foreach my $row (@multi_array) {
    foreach my $item (@$row) {
        $item =~ s/^\s+//;
    }
}
[download]