与价值不一致 [英] Inconsistency With the Value

问题描述

大家好，

在完全相同的公式行的代码中，我的C ++给了我一个答案，而excel给了另一个答案.
公式:

  double  val；

val = cos(- 0  .00623381826089208 + 90/180 *  3 . 12  .85 + cos(- 00623381826089208 )*(- 49 . 72 )+ 6767250 . 28012434 ; 

 c ++结果为6767200.5623347033，但excel结果显示为:6767200.641.因此，我将其拆分并比较了部分公式，
val = cos(-0.0062338182608920800000 + 90/180 * 3.14159265358979000);
 但同样，这在excel中显示0.006233778，在c ++中显示0.99998056981786743 

这在进一步的计算中产生了巨大的差异.我观察到的另一件事是行val = cos(20)或val = cos(10)给编译器带来了错误，并且cos只能取小数，例如0.2、0.9.我得到的错误是有多个实例的重载函数"cos"与参数列表匹配"

请需要帮助..

解决方案

您需要确保代码以正确的顺序进行评估，请参见C++中的运算符优先级 [double val ; val = cos(-0.00623381826089208+90/180*3.14159265358979)*12.85+cos(-0.00623381826089208)*(-49.72)+6767250.28012434;

The c++ result is 6767200.5623347033,  but the excel result shows:  6767200.641. SO I split it and compared a part of the formula,
val = cos(-0.0062338182608920800000+90/180 *3.14159265358979000);
 but again, this shows 0.006233778 in excel and 0.99998056981786743 in c++

This makes a huge difference in further calculations. Another thing I observed was the line val = cos(20) or val = cos(10) gives compiler error and cos can take only small numbers like 0.2, 0.9. The error I get is "more than one instance of overloaded function "cos" matches the argument list"

Need some help please..

解决方案

You need to make sure your code is evaluating in the correct order, see the rules of Operator precedence in C++[^] for guidance. You also need to check which version of the cos() function you are using, the compiler messages suggest it has a number of overloads. You may also find discrepancies owing to the way that floating point values are represented in computing and how accuracy can be a problem; a Google search will get you lots of articles on the subject.

I believe Excel considers argument of cos() function to be in radians while C++ treats it as degrees. Also, as Richard noted, check operator precedence in complex expressions like 90/180*3.14159265358979000, and make sure you divide float values, no ints

Watch the operator precedence and type casting rules! The term

90/180

delivers in C++ the value of 0, as it is an int division. Excel will probably convert both 90 and 180 to double first and comes to a result of 0.5.

Try writing 90./180. in your C++ expression and see the difference.

Also: Your expression is numerically not very stable. You arr adding something in the range of 50 to a value of 6767250. That''s 5 orders of magnitude difference. In type float you would probably get a totally erratic result. In double things run somewhat better, but don''t expect the result to have more than 15 - 5 = 10 significant digits.

这篇关于与价值不一致的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！