Using sprintf (problem with integer)

Hi to everybody, I have got a problem using sprintf. I'm using it for doing the following operation:

 step_value=sprintf('step%d',steps(i)*100);

..where "steps" is a simple array containing numbers with two decimals. The problem is that at i=9 I get:

steps(9)*100

ans = 55.0000

sprintf('%d',ans)

ans= 5.500000e+01

...which is not clearly in integer notation. I don't have problems for the other i-elements, just for i=9.

Someone knows what is it the problem? Thanks :)

Answer 1

[Walter Roberson]

http://matlab.wikia.com/wiki/FAQ#Why_is_0.3_-_0.2_-_0.1_.28or_similar.29_not_equal_to_zero.3F

Answer 2

[Andrea Bracchitta]

@Walter Roberson

I know this problem. Indeed, my variable steps=0.1:0.05:1.

However, before making the loop, I just write the following instruction:

steps=round(steps*100)/100;

In this way, all the "far" decimal numbers (due to floating point errors) should desappear and I should not have any problem. Am I saying something wrong in your opinion?

[Walter Roberson]

Imagine that you are using 2 digit decimal, starting from 5.33001 and you want to represent that as exact thirds.

5.33001 * 3 -> 15.99003
round(15.99003) -> 16
16/3 -> 5.3333333333333 ... infinitely in theory
5.3333333333333 to 2 decimal -> 5.33

and now you multiply that by 3

5.33 * 3 -> 15.99

Oops, you did not manage to find a value, X, such that X * 3 = 16 exactly.

This happens because 1/3 cannot be exactly represented in finite decimal.

The same difficulty happens in binary with respect to 1/10 : 1/10 cannot be represented exactly in finite binary. In binary it is .00011001100110011... When you take a finite leading portion of that and multiply it by 10, you get something that is not quite 1.