Casting between different datatypes in C++

Question

Working on casting between different datatypes in C++... The program here below prints:

>"Number is 2"
>"Number is 2.5"
>"Number is 2"

Please, explain why the last printout is not "Number is 2.5" which I would expect after the C++ style cast to float?

#include <iostream>
#include <conio.h>
using namespace std;

int main() {

    int iNumber = 5;
    float fNumber;

    // No casting - C++ implicitly converts the result into an int and saves into a float
    // which is a conversion from 'int' to 'float' with possible loss of data
    fNumber = iNumber / 2;
    cout << "Number is " << fNumber << endl;

    // C-style casting not recommended as not type safe
    fNumber = (float) iNumber / 2;
    cout << "Number is " << fNumber << endl;

    // C++ style casting using datatype constructors to make the casting safe
    fNumber = static_cast<float>(iNumber / 2);
    cout << "Number is " << fNumber << endl;

   _getch();

   return 0;
}

(iNumber / 2) becomes 2, which is then converted to a float, so the value will be 2. Order of operations. In the one before that, you cast iNumber to a float before it's divided, which is why it outputs 2.5. — Random Davis, Sep 27 '16 at 20:15

Jvinniec · Answer 1 · 2016-09-27T20:22:17.677

The statement

fNumber = static_cast<float>(iNumber / 2);

is dividing an integer by an integer BEFORE the cast to a float. This results in the following steps:

Divide int iNumber by int 2 --> results in an int of 2
Cast the result into a float --> results in a float of 2.

If you instead do:

fNumber = static_cast<float>(iNumber / 2.0);

Now the result of the division will be a floating type of 2.5 before the cast and you should get 2.5 as expected.

Thats all well and good, but then why does

fNumber = (float) iNumber / 2;

work? This is because you're casting iNumber to a float BEFORE the division operation, so here again, you're dividing a float by an int and the result will be a float of 2.5.

LastBlow · Answer 2 · 2016-10-07T15:41:31.057

static_cast<float>(expression) expects an expression in parentheses which parentheses will determine the order of the operations.

To effectively cast an int into a float without loss of data in the C++ style, the third expression needs to be rewritten with the parentheses around iNumber alone as in fNumber = static_cast<float>(iNumber) / 2; to eventually print "Number is 2.5":

iNumber is cast to float,
2 is implicitly cast to 2.0 and
iNumber is divided by 2.0.

Notice that the parentheses defining the order of operations are around the datatype in C but around the variable in C++.

Casting between different datatypes in C++

2 Answers2