about rounding off numbers

[flavourofbru]

Hi,

I have a small question.

If i need to round off a number to int, i do it like
int j = (int)(4.8);
thus "j" will have value 4;

but my requirement is to see that if the value is 4.8 then "j" should take value as 5 and if the value is 4.3 then "j" should take value as 4.

How can I do this in C++.

Thanks!!

[weaknessforcats]

Use a stringstream:

int Rounder(double d)
{
    stringstream str;
    str << d;
    int result;
    char decimalpoint;
    int decimal;
    str >> result >> decimalpoint >> decimal;
    if (decimal > 5)
    {
        ++result;
    }
    return result;
}
int main()
{
    double d1 = 4.8;
    cout << Rounder(d1) << endl;
    double d2 = 4.3;
    cout << Rounder(d2) << endl;
}
 

The notion of doing a cast is not rounding. That is covering up a compiler warning about truncation an possible loss of data.

The Rounder() function just reads the double into a stringstream and then extracts it as an int, char , int. Now you can make a test on the decmal portion and decide whether to increment the result or not.

Remember, when you cast in C++ you a) are calling a relic C function with a void* argument or b) your C++ design is faulty.

[flavourofbru]

Thanks for the clear explanation!!

[flavourofbru]

Hi,

the above is applicable only when i have a value between 4 and 5.
But in my program I am computing certain values and those values change each time in the loop. then in such case how can i round off as above??

do we have any in built functions for that??

Thanks!!

[r035198x]

Hi,

the above is applicable only when i have a value between 4 and 5.
...

No it is not limited to that range only.

[phiefer3]

Another option is to add 0.5 onto the number and then floor it, and then cast the result to an int.

if num is a double variable, then this would do it:

int j = static_cast<int>(floor(num + 0.5));

I believe you'll need to include <cmath> to use the floor function.

this works because the floor function always rounds down, and any number that with a decimal of .5 or higher is raised to the next integer before being rounded. so that with 4.3 is increased to 4.8 and then floored giving you 4; while 4.8 would be raised to 5.3 which is floored to 5.

This is assuming that you want a decimal value of .5 to round up, if you want .5 to round down, then you can do this:

int j = static_cast<int>(ceil(num - 0.5));

again needing to include <cmath> would be required.