ahoriuch has asked for the wisdom of the Perl Monks concerning the following question:

I'm having difficulties converting binary to double precision floating point. Using an on-line calculator, here is what the number should be: 5.71713084433028663131071134273E-5 Here's the number in hex: 

unpack("H16",$buffer);
#Returns 3f0d f967 3344 c570

[download]

Here is the number in "d": 

unpack("d",$buffer);
#Returns 1.69044404965372e+235

[download]

Something is clearly wrong. Reading the unpack documentation, there is no option for "D". Is there another way I can convert this number?

Replies are listed 'Best First'.
Re: Converting binary to double precision (big-endian)
by BrowserUk (Patriarch) on Feb 24, 2014 at 19:06 UTC

    Your value comes from a big-endian system, so you'll need 'd>' to decode it:

    $n = pack 'H16', '3f0df9673344c570';;

print unpack 'd>', $n;;
5.71713084433029e-005

    [download]

    You could also do the byte swapping yourself:

    $n = pack 'H16', '3f0df9673344c570';;

print unpack 'd', scalar reverse $n;;
5.71713084433029e-005

    [download]
    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.
[reply]
[d/l]
[select]

      In my case, this was the correct answer. I needed to switch from big to little-endian. It's amazing how little there is on the internet on this subject. Hopefully this will help future users. 

      my $n = pack 'H16', '3f0df9673344c570';;

printf "Big\n";
print unpack ("d", scalar reverse $n);

#Big
#5.71713084433029e-05

printf "\nLittle\n";
print unpack ("d",$n);
printf "\n";

#Little
#1.69044404965372e+235

      [download]
[reply]
[d/l]
Re: Converting binary to double precision
by toolic (Bishop) on Feb 24, 2014 at 18:25 UTC
    Reading the unpack documentation, there is no option for "D"
    pack 
    D  A float of long-double precision in native format.
   (Long doubles are available only if your system supports
   long double values _and_ if Perl has been compiled to
   support those. Raises an exception otherwise.)

    [download]
[reply]
[d/l]

      Is there a specific version/ install that I need? Invalid type 'D' in unpack at ... line 112, <FILE> chunk 1.

[reply]
        I'm running on 5.12.2, and I see it. What version of perl do you use?

        UPDATE: I don't see it in 5.8.8. Also, I can't find any mention of a change to pack in either perl5100delta or perl5120delta.

[reply]
Re: Converting binary to double precision
by Anonymous Monk on Feb 24, 2014 at 18:42 UTC
    If you know that you are dealing with an x86 platform, this approach might work, but Perl is platform-independent. The binary representation of a floating point number among CPU-types is not the same. For that matter, there is more than one floating-point type available on a single CPU. The binary representation will vary considerably.
[reply]

Back to Seekers of Perl Wisdom