What is the reason for this Perl formatting error?

Question

$str =sprintf "%014ld", "2555600000";
print $str;

This gives me the result:

-0001739367296

a negative number. Is there a reason for this behavior?

Answer 1

2555600000 is too big to fit in a 32-bit signed integer. Converted to binary, it's 10011000 01010011 01011100 10000000, which stands for -0001739367296 when interpreted as a signed 32-bit integer.

Your options are:

my $str = sprintf("%014u", 2555600000);     # unsigned integer
my $str = sprintf("%014.0f", 2555600000);   # floating point
my $str = sprintf("%014s", 2555600000);     # string

The first one works for this number, but not for much higher numbers. The second one can become inaccurate for very high numbers. (But with the standard floating point "double" you should easily be accurate for the 14 digits you're using. The third one might be the easiest, just treat it as a string and forget that it's a number.

To see how many bits your version is using, you can run perl -V. Example output:

...
intsize=4, longsize=4, ptrsize=4, doublesize=8, byteorder=1234
...

so my Perl is using 4-byte, 32-bit, integers.

Answer 2

$ perl -e '$str =sprintf "%014ld", "2555600000"; print $str, "\n";'
00002555600000

It works for me. However, the number you give is bigger than a 32 bit signed integer can hold, so maybe your result means you're on a 32 bit machine?