Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

BigDecimal multiply by zero

I am performing a simple multiplication with BigDecimal and I have found some strange behaviour when multiplying by zero (multiplying by zero is correct in this use-case).

Basic maths tells me that anything multiplied by zero will equal zero (see:Zero Product Property and Multiplication Properties)

However, the following code will consistently fail with the same error:

assertEquals(new BigDecimal(0), new BigDecimal(22.3).multiply(new BigDecimal(0))); 
java.lang.AssertionError:  Expected :0 Actual   :0E-48 

Is this an inaccuracy with BigDecimal or is there some niche branch of maths that I'm missing somewhere?

Notes: JDK 1.6.0_27 running in IntelliJ 11

like image 775
Richard Avatar asked Aug 14 '12 09:08

Richard


People also ask

How does BigDecimal value multiply?

Use the multiply() method to multiply one BigDecimal to another in Java. This method returns a BigDecimal whose value is (this × multiplicand), and whose scale is (this. scale() + multiplicand.

Can BigDecimal be double?

A BigDecimal is an exact way of representing numbers. A Double has a certain precision. Working with doubles of various magnitudes (say d1=1000.0 and d2=0.001 ) could result in the 0.001 being dropped alltogether when summing as the difference in magnitude is so large. With BigDecimal this would not happen.

Does BigDecimal accept negative values?

The value of the BigDecimal is (BigInteger/10**scale). A negative scale will result in a NumberFormatException.


2 Answers

You can't use the equals() method to compare BigDecimals, like this assertion does. That is because this equals function will compare the scale. If the scale is different, equals() will return false, even if they are the same number mathematically.

You can however use compareTo() to do what you want:

As @assylias points out, you should also use the new BigDecimal("22.3") constructor to avoid double precision issues.

BigDecimal expected = BigDecimal.ZERO; BigDecimal actual = new BigDecimal("22.3").multiply(BigDecimal.ZERO); assertEquals(0, expected.compareTo(actual)); 

There is also a method called signum(), that returns -1, 0 or 1 for negative, zero, and positive. So you can also test for zero with

assertEquals(0, actual.signum()); 
like image 78
Keppil Avatar answered Sep 22 '22 18:09

Keppil


There are 2 issues with your code:

  • you should compare BigDecimal with compareTo instead of equals, as advised by the other answers
  • but you should also use the string constructor: new BigDecimal("22.3") instead of the double constructor new BigDecimal(22.3) to avoid double precision issues

In other words, the following code (which correctly uses compareTo) still returns false:

BigDecimal bd = new BigDecimal(0.1).multiply(new BigDecimal(10)); System.out.println(bd.compareTo(BigDecimal.ONE) == 0); 

because 0.1d * 10d != 1

like image 31
assylias Avatar answered Sep 23 '22 18:09

assylias