Why does Resharper complain when I compare a double to zero?

问题

If I do

double d = 0;
if (d == 0) {
  ...
}

Resharper complains at the comparison d == 0 about "Comparison of floating point number with equality operator. Possible loss of precision while rounding values."

Why? It cannot be difficult to represent exact zero as a double or a float can it?

I understand such a warning would be relevant if I compared to some other value such as 0.2 for which there is no exact binary representation.

回答1:

Resharper does not analyze how the double variable got its value.

After a few calculations a double value is rarely exact so resharper warns you that comparing a double with an exact value is not a good idea.

double x = Math.Sqrt(2);
double d = x * x;

Console.WriteLine(d == 2);

回答2:

Since R# 6, many such inspections have a 'Why is ReSharper suggesting this?' item on their Alt+Enter menu. In this case, the explanation relates to the possible unintended consequences of doing equality comparisons on floating point values:

Using the == operator to compare floating-point numbers is, generally, a bad idea. The problem arises from the fact that, typically, results of calculations have to be ‘fitted’ into floating-point representation, which does not always matched the perceived reality of what result should be produced.

回答3:

often calculation with double is inexact. comparing a double with an exact value may be problematic. Comparing with an intervallmight be more secure.

if ((d > -0.000001) && (d < +0.000001)) {
   ...
}

the same appy when comparing dates

if ((date >= DateTime.parse("2012-05-21T00:00:00")) && 
   (date <= DateTime.parse("2012-05-21T23:59:59"))) {
}

来源：https://stackoverflow.com/questions/10684365/why-does-resharper-complain-when-i-compare-a-double-to-zero

标签

floating-point

resharper