Win a copy of Mesos in Action this week in the Cloud/Virtualizaton forum!

# Arithmetic operations with large numbers

Devi Hari
Greenhorn
Posts: 7
How can i do arithmetic operations with large numbers in java?
when i use double, it returns the result with E, can i avoid this way of representation?
simple multiplication like 154.2 * 100000 giving result as 1.5419999999999998E7

Ulf Dittmer
Rancher
Posts: 42967
73
Java has two classes for working with large integer and floating-point numbers. They're called java.math.BigInteger and java.math.BigDecimal, respectively.

Devi Hari
Greenhorn
Posts: 7
Originally posted by Ulf Dittmer:
Java has two classes for working with large integer and floating-point numbers. They're called java.math.BigInteger and java.math.BigDecimal, respectively.

I've already tried using BigDecimal, but got problems with precision

this is the code :

double dAmnt = 154.20;
double amt = 1000000000;

BigDecimal bd1 = new BigDecimal(dAmnt);
BigDecimal bd2 = new BigDecimal(amt);
BigDecimal bd3 = bd1.multiply(bd2);
System.out.println("Bigdecimal == "+bd3);

I am getting the result as 154199999999.99998863131622783839702606201171875000000000

Peter Chase
Ranch Hand
Posts: 1970
The problem is that you are constructing your BigDecimal from a "double". But the numbers you are trying to use cannot be perfectly represented by a double, so are wrong before ever becomeing a BigDecimal. To confirm this, print out the values of your BigDecimals, before the multiplication.

Instead, you should avoid floating-point (e.g. "double") altogether and construct your BigDecimal from a String or from a BigInteger. See the API for details of these constructors.

Any attempt to use floating-point arithmetic in combination with BigDecimal or BigInteger is likely to end in disappointment.

Devi Hari
Greenhorn
Posts: 7
Originally posted by Peter Chase:
The problem is that you are constructing your BigDecimal from a "double". But the numbers you are trying to use cannot be perfectly represented by a double, so are wrong before ever becomeing a BigDecimal. To confirm this, print out the values of your BigDecimals, before the multiplication.

Instead, you should avoid floating-point (e.g. "double") altogether and construct your BigDecimal from a String or from a BigInteger. See the API for details of these constructors.

Any attempt to use floating-point arithmetic in combination with BigDecimal or BigInteger is likely to end in disappointment.

That solved the problem.. Thanks a lot!!!