CODEWITHSUNDEEP
c++visual-c++
Abhinav Konda
Abhinav Konda

Reputation: 225

Interpretation error of double data type

I read a double value using std::cin from the keyboard, let the value be 1.15 . When I place a break point after reading the value visual studio showed that value as 1.14999999. But If I print it It printed 1.15 on my console. Later I wrote following code and it did not work well

int main()
{
    long double valueA;
    int required;
    std::cin>>valueA;
    required=(valueA*10000)-(((int)valueA)*10000);
    std::cout<<(required);
}

When the input is 1.015 the output is 149 but the expected output is 150. Why is my compiler considering 1.015 as 1.014999999? How can I correct that error?

Upvotes: 1

Views: 55

Answers (1)

Radu Chivu
Radu Chivu

Reputation: 1065

What you are describing is floating point error. This happens because of the way floating point is represented at the hardware level (see here). Basically a floating point number is kept as sme and is reconstructed as s * m * 2 ^ e where ^ is to the power of and s is 1 if the s bit is 0 and -1 if the s bit is 1.

If you need the sort of accuracy, you can use a decimal arithmetic library, it's more or less the same thing, but instead of using powers of 2, they use powers of 10 and because they are implemented in software, it means they can have arbitrary precision (more on that here).

Here's a list of libraries that implement decimal arithmetic that you can use:

Upvotes: 2

Related Questions