in reply to How to determine whether a number has a decimal component?

Because of limitations on integer vs floating point representations in computing, comparing $number vs int($number) may cause problems with integer overflow (use bigint may alleviate this). For input data, checking for a decimal point should work; for the results of computation, well, remember that on most computers (1.0/3.0) * 3.0 won't return exactly 1, but something like 0.99999…, so, as mentioned by others, you'll have to check for $number - int($number) ≤ &epsilon. You'll also have to realize that it won't always work.

Information about American English usage here and here. Floating point issues? Please read this before posting. — emc

  • Comment on Re: How to determine whether a number has a decimal component?

Replies are listed 'Best First'.
Re^2: How to determine whether a number has a decimal component?
by Xenofur (Monk) on Jan 04, 2009 at 18:38 UTC
    Does this make it possible to get false negatives, i.e. $x is 0.999 and comparing it against int $x gives the result that they ARE equal? If not, this is of no interest to me at all, since false positives would not have any adverse impact if they only result into action that ensures that $x does not contain a decimal component.
[reply]

      Well, that particular case won't unless the tolerance (ε) is very large, as int(0.999) is 0. ;-) More generally, it is, of course, possible, although when it happens will be dependent on the details of the C compiler and compiler options used to build a particular installation of Perl. Alas, this means that it's possible that different Perl installations on a given platform, e.g., ActiveState and Strawberry Perl on Windows, could yield different results.

      Given your last statement, "false positives would not have any adverse impact if they only result into action that ensures that $x does not contain a decimal component." you may want to tell us why you're interested in this test, at all. It may be superfluous.

      Information about American English usage here and here. Floating point issues? Please read this before posting. — emc

[reply]
        Ok, i'll try to give a simple explanation. I was drawing a graph with GD::Graph and would like to restrict the the y-ticks so they only get drawn on values like these: 1, 1.5, 2, 2.5, etc. The number of ticks was pre-set.

        I've since gone another way that dispenses with this completely, but still thought it would be something nice to meditate over, especially as i couldn't find a solution that didn't fall afoul any of these:
        • shooting with cannons at flies
        • incredibly hackish
        • requires a lot of code
        • comes with a lot of situations where it fails
        I was hoping i'd missed something and someone would point out some inbuilt "has_decimal" function in Perl or something, but apparently no such thing exists.
[reply]

In Section Seekers of Perl Wisdom