The Class BigDecimal

Imprecision in calculations

In the Apples by the Box problem, we wrote a program that computes the money saved by buying apples by the box instead of individually. For a box of 24 apples, the savings should be the difference between $7.68 and $7.25, or $0.43. Instead of displaying $0.43, the program displayed $0.4299999999999997. This imprecision is the result of using the binary number system to represent decimal values. Although we can force the output to look like $0.43, the internal representation of the result is not quite correct. If we used this value in subsequent computations, the error would propagate and likely get larger. Using binary arithmetic for monetary computations is not a good idea, and is certainly something that banks could not tolerate.

Accurate decimal arithmetic

The class BigDecimal enables us to perform accurate decimal arithmetic and to work with arbitrarily large numbers. This class is in the package java.math. Do not confuse this package with the class Math which is in the package java.lang.

✏️ Programming tip:

To use the class BigDecimal in your program, be sure to precede your class with the following statement:

 import java.math.BigDecimal;

📝 Note:

An object of BigDecimal has a signed decimal value and a scale, both of which are integers. The value of the scale positions the decimal point within the value. Usually, this representation need not concern you, but the terms value and scale do appear within the description of some of the class’s methods.

Creating a BigDecimal object

To create a new BigDecimal object, we use:

BigDecimal price = new BigDecimal("2.97");

We write the desired value of the new object as a string to preserve its accuracy. Writing it as a double is legal, but doing so would involve a conversion to binary and, therefore, an inaccuracy in its representation.

If we want to read the value of the BigDecimal object from the user, we could read it as a string. For example: