I am testing one of my methods, double getSqrt(double s)
, and having the following JUnit test code:
我正在测试我的一个方法,双getSqrt(double s),并具有以下JUnit测试代码:
@Test
public void testGetSqrt() {
System.out.println("getSqrt");
double s = 16.0;
Calculator instance = new Calculator();
double expResult = 4.0;
double result = 4.000000000052429; // was "instance.getSqrt(s);" here
// now hardcode for test purpose
System.out.println(System.getProperty("java.class.path"));
assertEquals(expResult, result, 0.0);
}
It failed every time, no matter how I changed the tolerance "0.0", "0.1", "0.00", "0.01", "0.001".... The report is the following:
无论我如何改变公差“0.0”,“0.1”,“0.00”,“0.01”,“0.001”......它每次都失败了....报告如下:
AssertionFailedError: expected:<4.0>; but was:<4.000000000052429>
AssertionFailedError:expected:<4.0>;但是:<4.000000000052429>
What I am using is JUnit 4.12 and Hamcrest 1.3.
我使用的是JUnit 4.12和Hamcrest 1.3。
And my imports:
我的进口:
import org.junit.Test;
import static org.junit.Assert.*;
Did I use the tolerance wrong?
我使用了容差吗?
Output on the console:
控制台上的输出:
compile-test-single:
Testsuite: descriptivestatisticsapplication.processing.CalculatorTest
getSqrt
Tests run: 1, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 0.079 sec
------------- Standard Output ---------------
getSqrt
C:\Java\jdk1.8.0_111\jre\lib\javaws.jar;C:\Java\jdk1.8.0_111\jre\lib\deploy.jar;C:\Java\jdk1.8.0_111\jre\lib\plugin.jar;C:\Users\Z\Documents\NetBeansProjects\DescriptiveStatisticsApplication\build\classes;C:\Program Files\NetBeans 8.2\platform\modules\ext\junit-4.12.jar;C:\Program Files\NetBeans 8.2\platform\modules\ext\hamcrest-core-1.3.jar;C:\Users\Z\Documents\NetBeansProjects\DescriptiveStatisticsApplication\build\test\classes;C:\Program Files\NetBeans 8.2\extide\ant\lib\ant-launcher.jar;C:\Program Files\NetBeans 8.2\extide\ant\lib\ant.jar;C:\Program Files\NetBeans 8.2\extide\ant\lib\ant-junit.jar;C:\Program Files\NetBeans 8.2\extide\ant\lib\ant-junit4.jar
------------- ---------------- ---------------
Update on 5/27:
5月27日更新:
After doing some tests from the suggestions in the comments, this unit test can now pass successfully. I think this question can be closed.
在根据评论中的建议进行一些测试后,此单元测试现在可以成功通过。我认为这个问题可以结束。
What I worry about is I didn't find out any reason why it failed and why now it succeed! Same code, same logic! The only thing I have done is like commenting out some non-important lines and then adding them back, changing some declaration statements sequence and then changing them back. The same code has different behaviors. I haven't changed anything in the building and running env and libs. This doesn't make any sense. The result from a piece of code should be reproducible.
我担心的是我没有找到任何失败的原因,为什么现在它成功了!相同的代码,相同的逻辑!我唯一做的就是注释掉一些非重要的行然后再添加它们,更改一些声明语句序列然后再更改它们。相同的代码有不同的行为。我没有在构建和运行env和libs中改变任何东西。这没有任何意义。一段代码的结果应该是可重现的。
1 个解决方案
#1
-1
The float in Java is not precise and as you've already seen the result <4.000000000052429> In other words, you set a float number to 4 in java, it is not actually 4. You can use long or even BigInteger to solve this problem.
Java中的浮点数并不准确,因为您已经看到结果<4.000000000052429>换句话说,您在java中将浮点数设置为4,实际上不是4.您可以使用long或甚至BigInteger来解决此问题。
#1
-1
The float in Java is not precise and as you've already seen the result <4.000000000052429> In other words, you set a float number to 4 in java, it is not actually 4. You can use long or even BigInteger to solve this problem.
Java中的浮点数并不准确,因为您已经看到结果<4.000000000052429>换句话说,您在java中将浮点数设置为4,实际上不是4.您可以使用long或甚至BigInteger来解决此问题。