Problem with calculating floats

strange situation, when performing the following lines of Code:

const float a = 47.848711;
const float b = 47.862952;
float result = b - a;

I get a (NSLog %.10f) result = 0.0142440796.

I expect开发者_开发问答ed to get 0.0142410000.

What's going on?

Classic!

What Every Computer Scientist Should Know About Floating-Point Arithmetic

(basically, floating points can be inaccurate; wikipedia).

What if I ask you the following:

const int a = 1.3;
const int b = 2.7;
int result = b - a;

I get a (NSLog %d) result = 1.

I expected to get 1.4. What's going on?

In this case, the answer is obvious, right? 1.3 isn't an integer, so the actual value that gets stored in a is 1, and the value that gets stored in b isn't 2.7, but rather 2. When I subtract 1 from 2 I get exactly 1, which is the observed answer. If you're with me so far, keep reading.

The exact same thing is happening in your example. 47.848711 isn't a single-precision float, so the closest floating-point value is stored in a instead, which is exactly:

a = 47.8487091064453125

Similarly, the value stored in b is the closest floating-point value to 47.862952, which is exactly:

b = 47.86295318603515625

When you subtract these numbers to get result, you get:

  47.86295318603515625
- 47.8487091064453125
----------------------
   0.01424407958984375

When you round that value to 10 digits to print it out, you get:

   0.0142440796