Java中的float、double计算精度问题
java中的float、double计算存在精度问题,这不仅仅在java会出现,在其他语言中也会存在,其原因是出在IEEE 754标准上。
而java对此提供了一个用于浮点型计算的类——BigDecimal(java.math.BigDecimal),通过将double替换成BigDecimal进行计算可以获得较为精确的计算结果。
BigDecimal的构造方法有许多,在此推荐使用BigDecimal(String val)的构造方法,通过String字符串进行构造。可能会有人直接使用BigDecimal(double val)去构造,但为什么推荐要使用String而不用double直接构造?原因如BigDecimal(double val)前的注释所说:
Translates a {@code double} into a {@code BigDecimal} which is the exact decimal representation of the {@code double}'s binary floating-point value. The scale of the returned {@code BigDecimal} is the smallest value such that (10 scale val) is an integer.
Notes: The results of this constructor can be somewhat unpredictable. One might assume that writing {@code new BigDecimal(0.1)} in Java creates a {@code BigDecimal} which is exactly equal to 0.1 (an unscaled value of 1, with a scale of 1), but it is actually equal to 0.1000000000000000055511151231257827021181583404541015625.This is because 0.1 cannot be represented exactly as a {@code double} (or, for that matter, as a binary fraction of any finite length). Thus, the value that is being passed in to the constructor is not exactly equal to 0.1, appearances notwithstanding.
The {@code String} constructor, on the other hand, is perfectly predictable: writing {@code new BigDecimal("0.1")} creates a {@code BigDecimal} which is exactly equal to 0.1, as one would expect. Therefore, it is generally recommended that the {@linkplain #BigDecimal(String) String constructor} be used in preference to this one.
When a {@code double} must be used as a source for a {@code BigDecimal}, note that this constructor provides an exact conversion; it does not give the same result as converting the {@code double} to a {@code String} using the {@link Double#toString(double)} method and then using the {@link #BigDecimal(String)} constructor. To get that result, use the {@code static} {@link #valueOf(double)} method.
balabala的一大段?挑出重点看,Notes(说明):The results of this constructor can be somewhat unpredictable(这个结果可能是不可预测的)。原因在后面一句话也说了,若用 0.1 new一个BigDecimal,则它实际上是等于0.1000000000000000055511151231257827021181583404541015625,原因无需细看,跳过。
The {@code String} constructor, on the other hand, is perfectly predictable(通过字符串构造,是可以完全预测的)。到此注释说的差不多了。
而BigDecimal中一系列的add、subtract等方法对应着加减乘除就不必多说。
到这里就结束了吗?不,如果单纯的将double替换成BigDecimal,就会大幅降低程序的运行速度,因此需要进行一定的优化:非替换,而是改进。
依旧用double定义、储存数据,但计算时,使用BigDecimal进行计算(若需要精确的计算),最后只需.doubleValue()即可得到较为精确的double类型的计算结果了。
由此可以编写一个用于浮点型计算的工具类,专职浮点型计算工作。
1 import java.math.BigDecimal; 2 3 /** 4 * 浮点型较为精确计算工具类 5 */ 6 public class CalculateUtil { 7 /** 8 * 私有构造 9 */ 10 private CalculateUtil(){} 11 /** 12 * 加法运算 13 * @param num1 被加数 14 * @param num2 加数 15 * @return 两数之和 16 */ 17 public static double add(double num1, double num2){ 18 BigDecimal b1 = new BigDecimal(Double.toString(num1)); 19 BigDecimal b2 = new BigDecimal(Double.toString(num2)); 20 return b1.add(b2).doubleValue(); 21 } 22 /** 23 * 减法运算 24 * @param num1 被减数 25 * @param num2 减数 26 * @return 两数之差 27 */ 28 public static double sub(double num1, double num2){ 29 BigDecimal b1 = new BigDecimal(Double.toString(num1)); 30 BigDecimal b2 = new BigDecimal(Double.toString(num2)); 31 return b1.subtract(b2).doubleValue(); 32 } 33 /** 34 * 乘法运算 35 * @param num1 被乘数 36 * @param num2 乘数 37 * @return 两数之积 38 */ 39 public static double mul(double num1, double num2){ 40 BigDecimal b1 = new BigDecimal(Double.toString(num1)); 41 BigDecimal b2 = new BigDecimal(Double.toString(num2)); 42 return b1.multiply(b2).doubleValue(); 43 } 44 /** 45 * 除法运算(小数点后10位) 46 * @param num1 被除数 47 * @param num2 除数 48 * @return 两数之商 49 */ 50 public static double div(double num1, double num2){ 51 return div(num1, num2, 10); 52 } 53 54 /** 55 * 除法运算 56 * @param num1 被除数 57 * @param num2 除数 58 * @param scale 小数点后精度位数 59 * @return 两数之商 60 */ 61 public static double div(double num1, double num2, int scale){ 62 if(scale<0) 63 throw new IllegalArgumentException("The scale must be a positive integer or zero"); 64 BigDecimal b1 = new BigDecimal(Double.toString(num1)); 65 BigDecimal b2 = new BigDecimal(Double.toString(num2)); 66 return b1.divide(b2, scale, BigDecimal.ROUND_HALF_UP).doubleValue(); 67 } 68 /** 69 * 四舍五入 70 * @param num 需要四舍五入的数 71 * @param scale 小数点后精度位数 72 * @return 四舍五入值 73 */ 74 public static double round(double num, int scale){ 75 if(scale<0) 76 throw new IllegalArgumentException("The scale must be a positive integer or zero"); 77 BigDecimal b1 = new BigDecimal(Double.toString(num)); 78 return b1.divide(new BigDecimal("1"), scale, BigDecimal.ROUND_HALF_UP).doubleValue(); 79 } 80 81 }