Why is this floating point comparison true?

Question

Possible Duplicate:
Floating point comparison

When I run the code below I get the following output:

Output : if

Why does this happen?

#include <stdio.h>

void main()
{
    float a = 0.7;
    if(a<0.7)
    {
        printf("if");
    }
    else
    {
        printf("Else");
    }
}

[What every programmer should know about floating-point arithmetic](http://floating-point-gui.de/) — Bart, Oct 31 '11 at 08:18
@downvoter dont be panic just by asking same question by new user ...close such questan as duplicate but dont downvote it — Jeegar Patel, Oct 31 '11 at 08:26
This an applet which illustrates the internal working a bit better: http://www.h-schmidt.net/FloatApplet/IEEE754.html — ChrisWue, Oct 31 '11 at 08:28

Pascal Cuoq · Answer 1 · 2011-10-31T08:31:08.410

Your program compares the double representation of 0.7 (which is what the compiler produces when you write 0.7 in your program) with a float representation of 0.7 (which is what variable a contains). The two are not the same, and it just happens that the double one is slightly larger.

EDIT: (float)0.7 can be represented compactly in hexadecimal notation. It is 0x1.666666p-1. 0.7 as a double constant is 0x1.6666666666666p-1, so it is slightly larger.

score 2 · Accepted Answer · edited Oct 31 '11 at 08:26

2

Floating points are not stored in precise format. Most likely, your platform interprets

float a = 0.7;

as

float a = 0.69999....;

This is because of the internal representation of floating points on your platform. The link provided by Daniel should get you started.

edited Oct 31 '11 at 08:26

Williham Totland

27,735
5
49
68

answered Oct 31 '11 at 08:23

Luchian Grigore

245,575
61
446
609

2

this still doesn't answer why a < 0.7 (see link in the comments) – Karoly Horvath Oct 31 '11 at 08:32

score 0 · Answer 3 · answered Oct 31 '11 at 08:19

0

That is due to accuracy issues in floating point representation on a computer. See this Wikipedia article.

answered Oct 31 '11 at 08:19

Daniel Rose

16,753
9
65
86

Why is this floating point comparison true?

3 Answers3