$lnpam[$i][$j] = log $matrix[$i][$j] if $matrix[$i][$j];
[download]

The resulting matrix would have undefined values in the corresponding places.

i solved the problem using this:

LINE: for (my $i=0; $i<20; $i++){
    for (my $j=0; $j<20; $j++){
        if ($matrix[$i][$j] =~ /[^0.00]/){
            $lnpam[$i][$j] = log $matrix[$i][$j];
        }else{ next LINE;
        }
    }
}
[download]

thanks a lot for the quick answer!

if ($matrix[$i][$j] =~ /[^0.00]/){

This does not do what you think it does, but it is at least close. You are really only doing if ($matrix[$i][$j] =~ /[^0.]/){ IE: does the string have at least one character that is not a literal '0' or a '.'

See: Bracketed Character Classes

Why not make the test if ($matrix[$i][$j] > 0)? That's much simpler and faster, and will not fail on "-0.00 " or " 0.00" or regular negative numbers.

Rather than having to testing for 0 (or undef) in all your further calculations, it may be easier (and mathematically acceptable) to replace your zeros with a very small number, say 1e-308:

map $_//= 1e-308, @$_ for @matrix;
[download]

which should have negligible affect upon any calculations, but avoid lot of special casing.

The POSIX HUGE_VAL would be closer to the mathamatical concept. It would never be mistaken for "unassigned". Attempts to use it in an expression would always throw an exception. A huge negative would be even better, but we do not seem to have that option.
Bill

Attempts to use it in an expression would always throw an exception.

I think you've misunderstood my post. If I wanted to causes exceptions, I'd leave the zeros in place. or replace them with #NaN.

My point was that by replacing the zeros with a very, very small number, you can avoid exceptions, or the need for conditional tests to avoid exceptions, as it will allow most calculations to operate correctly whilst introducing negligible changes to the final results.

---

With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'

Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.

"Science is about questioning the status quo. Questioning authority".

In the absence of evidence, opinion is indistinguishable from prejudice.

HUGE_VAL does not help here, we need kind of a TINY_VAL. 1 / HUGE_VAL does not work:

$ perl -MPOSIX=HUGE_VAL -e 'print log(1 / HUGE_VAL)'
Can't take log of 0 at -e line 1.
[download]

لսႽ† ᥲᥒ⚪⟊Ⴙᘓᖇ Ꮅᘓᖇ⎱ Ⴙᥲ𝇋ƙᘓᖇ

You have several 0.00 entries in your example. What do you want to output for log(0)?

Thanks you all for the reply! it really helped!

My suggestion: don't try to take the log of zero as it's not defined :) see logarithm

Really, it'll depends on what you're going to do with the logs. You could set the result to undef and detect that in your later code.

Computationally hazardous(!), but you could try one of these modifications with e.g. 1e-300 as an approximation for zero and see how far you get... Take care of error propagation.

  $lnpam[$i][$j] = log( $matrix[$i][$j]  || 1e-300);
[download]

  $lnpam[$i][$j] = $matrix[$i][$j]  ? log( $matrix[$i][$j]) : -1000;
[download]