DrLudwig3
DrLudwig3

Reputation: 421

Android - exact mathematical calculation

I have got a Problem, I am developing an Application which should be able to do some mathematic calculations. These calculations have to be exact (or rather not obviously wrong)

But this simple Code

double a = 3.048d;
double b = 1000d;

double c = a / b;

gives me a wrong result c is not 0.003048 as expected instead it is 0.0030480000000000004 which is obviously wrong.

double d = 3.048 / 1000; 

this second code-snipet gives the correct result.

I am aware that all floatingpoint arithmetic is not exact when calculating with computers but I don't know how to solve this problem.

thanks in advance!
Ludwig

Developing for:
- Android 2.2
Testdevice:
- HTC Desire

Upvotes: 0

Views: 395

Answers (4)

Use a BigDecimal for precise floating point calculations. Setting the scale allows you to specify precisely how far out you want to go for output.

import java.math.BigDecimal;

class Test{

        public static void main(String[] args){
                BigDecimal a = new BigDecimal("3.048");
                BigDecimal b = new BigDecimal(1000);
                BigDecimal c = a.divide(b).setScale(6);
                System.out.println(c); //0.003048
        }
}

Upvotes: 1

jimw
jimw

Reputation: 2598

This is a consequence of the IEEE 754 floating point representation, not an error. To deal with it, round your result to an appropriate precision.

Upvotes: 1

John Ericksen
John Ericksen

Reputation: 11113

What you need to use for exact percision is the BigDecimal object:

BigDecimal a = new BigDecimal("3.048");
BigDecimal b = new BigDecimal(1000);

BigDecimal c = a.divide(b);

System.out.println(c); //0.003048

Upvotes: 5

Christopher Oezbek
Christopher Oezbek

Reputation: 26333

Use BigDecimal for such precise allocations.

Btw the result for d is obviously right, because double has a machine encoding which cannot store the result, which you perceive as correct.

Upvotes: 0

Related Questions