how to write a function to round a float

Really,if I write round(4.015,2),it returns 4.01,but I hope 4.02.

Then, can use bitwise operation to slove this? or any other ways to slove?

OxBADC0DE (240)

Use iomanip.h and use this:
cout<<setprecision(2)<<fixed<<showpoint<<x<<endl;
where your x is the number to round up and 2 is the number to round up to

I don't need output it. I just want to return a value to other operations.

I guess your 4.015 is actually already rounded from 4.0125. Give it a try with 4.0124 and 4.0126. One might return 4.01 and the other 4.02.

Last edited on

#include <iostream>
using namespace std;

double round_number(double, int);
int pow_new(int, int number = 10);

int main ()
{
	double number;
	int places;
	
	go_back:
	
	cout << "Enter double number to round: ";
	cin >> number;
	cout << "Enter places to round: ";
	cin >> places;
	
	cout << round_number(number, places) << endl;
	
	goto go_back;
	
	return 0;
}

double round_number(double number, int places)
{
	places = pow_new(places);
	int temp_number = number * places;
	double number_2 = (double)temp_number / (double)places;
	double number_3 = (number - number_2) * places * 10.0;
	if (number_3 >= 5)
	{
		number_2 = number_2 + (1.0 / places);
	}
	return number_2;
}

int pow_new(int places, int number)
{
	if (places == 0)
	{
		return 1;
	}
	else if (places == 1)
	{
		return number;
	}
	else
	{
		return pow_new(places - 1, number * 10);
	}
}

Last edited on

Your header says something about float but none of the codes above has any. Is the problem about rounding floats or just rounding?

Lowest0ne (1536)

double round(double d, int pp) // pow() doesn't work with unsigned, so made this switch.
{ 
	return int(d * pow(10.0, pp) + .5) /  pow(10.0, pp);
}

Last edited on

Sorry for that one. :)

Thanks for help!
The code above can work well.
But some examples are not which I expect.
Such as:
number is 4.015,places are 2, I expect to get 4.02,but it is 4.01;
number is 4.2225,places are 3,I expect to get 4.223,but it is 4.222

Is this because float number can't be very precision?
Do I have any ways to slove this?

That's what your exercise is about. You cannot solve this. The reason is the binary. 1/2 + 1/4 + 1/8 + ...

Last edited on

Stewbond (2827)

another useful tool is fmod() from <cmath>.
http://www.cplusplus.com/reference/clibrary/cmath/fmod/

int ftoi(float f)
{
    return (int)f + ((fmod(f,1.f) < 0.5f)? 0 : 1);
}

You can play around with it for different precisions

Last edited on

solution

double round_number(double number, int places)
{
	return int(number * pow_new(places) + 0.50001) / double(pow_new(places));
}

if places are no more then 5.

Thanks!
if places are no more then 5,you advice is helpful.
But I have an another occasion:

number is 4.499999,places are 0, I expect to get 4,but it is 5
I try to subtract a very small number,but not succeed.
How should I write?

write return int(number * pow_new(places) + 0.5000001) / double(pow_new(places));
but you will get 5 if number is 4.49999999, places are 0. :D
in other words:
if you write 0.5001 number 4.4999, places 0 will be 5.
if you write 0.50001 number 4.49999, places 0 will be 5
and so on.

or
return int(number * pow_new(places) + 0.5 + 1e-16) / double(pow_new(places));
16 is your precision. On my system this is max precision.
15 places after .

Last edited on

Thanks!
but what if I convert 4.49999999 to a string to choose how to round?

You can round like that too.

I look again and found more bugs in
return int(number * pow(10.0, places) + 0.501) / double(pow(10.0, places));

Enter double number to round: 4.3
Enter places to round: 8
4.3
Enter double number to round: 4.3
Enter places to round: 9
-2.14748
Enter double number to round: 4.3
Enter places to round: 10
-1.52297
Enter double number to round: 4.3
Enter places to round: 11
-1.76638
Enter double number to round: 4.3
Enter places to round: 12
2.95235
Enter double number to round: 4.3
Enter places to round: 13
-1.63166
Enter double number to round: 4.3
Enter places to round: 14
4.3
Enter double number to round: 4.3
Enter places to round: 15
1.40313

Last edited on