我如何用c#优雅地做到这一点?
例如,一个数字可以是1到100之间。
我知道一个简单的if (x >= 1 && x <= 100)就足够了;但是有很多语法糖和新特性不断添加到c# /。Net这个问题是关于更习惯的(一个可以称之为优雅的)写法。
性能不是问题,但请在非O(1)的解决方案中添加性能说明,因为人们可能会复制粘贴建议。
我如何用c#优雅地做到这一点?
例如,一个数字可以是1到100之间。
我知道一个简单的if (x >= 1 && x <= 100)就足够了;但是有很多语法糖和新特性不断添加到c# /。Net这个问题是关于更习惯的(一个可以称之为优雅的)写法。
性能不是问题,但请在非O(1)的解决方案中添加性能说明,因为人们可能会复制粘贴建议。
当前回答
当检查一个“数字”是否在一个范围内时,你必须清楚你的意思,两个数字相等意味着什么?一般来说,你应该把所有浮点数包装在一个所谓的“epsilon球”中,这是通过选择一个小的值来完成的,如果两个值如此接近,它们就是相同的。
private double _epsilon = 10E-9;
/// <summary>
/// Checks if the distance between two doubles is within an epsilon.
/// In general this should be used for determining equality between doubles.
/// </summary>
/// <param name="x0">The orgin of intrest</param>
/// <param name="x"> The point of intrest</param>
/// <param name="epsilon">The minimum distance between the points</param>
/// <returns>Returns true iff x in (x0-epsilon, x0+epsilon)</returns>
public static bool IsInNeghborhood(double x0, double x, double epsilon) => Abs(x0 - x) < epsilon;
public static bool AreEqual(double v0, double v1) => IsInNeghborhood(v0, v1, _epsilon);
有了这两个辅助,并假设任何数字都可以转换为double而不需要所需的精度。现在需要的是一个枚举和另一个方法
public enum BoundType
{
Open,
Closed,
OpenClosed,
ClosedOpen
}
另一种方法如下:
public static bool InRange(double value, double upperBound, double lowerBound, BoundType bound = BoundType.Open)
{
bool inside = value < upperBound && value > lowerBound;
switch (bound)
{
case BoundType.Open:
return inside;
case BoundType.Closed:
return inside || AreEqual(value, upperBound) || AreEqual(value, lowerBound);
case BoundType.OpenClosed:
return inside || AreEqual(value, upperBound);
case BoundType.ClosedOpen:
return inside || AreEqual(value, lowerBound);
default:
throw new System.NotImplementedException("You forgot to do something");
}
}
现在,这可能远远超过了您想要的,但它使您不必一直处理舍入问题,并试图记住一个值是否被舍入到哪个位置。如果你需要,你可以很容易地将它扩展到任意的情况并允许变化。
其他回答
只是为了增加这里的噪音,你可以创建一个扩展方法:
public static bool IsWithin(this int value, int minimum, int maximum)
{
return value >= minimum && value <= maximum;
}
这样你就能做…
int val = 15;
bool foo = val.IsWithin(5,20);
话虽如此,当检查本身只有一行时,这样做似乎是一件愚蠢的事情。
In C, if time efficiency is crucial and integer overflows will wrap, one could do if ((unsigned)(value-min) <= (max-min)) .... If 'max' and 'min' are independent variables, the extra subtraction for (max-min) will waste time, but if that expression can be precomputed at compile time, or if it can be computed once at run-time to test many numbers against the same range, the above expression may be computed efficiently even in the case where the value is within range (if a large fraction of values will be below the valid range, it may be faster to use if ((value >= min) && (value <= max)) ... because it will exit early if value is less than min).
不过,在使用这样的实现之前,请先对目标机器进行基准测试。在某些处理器上,由两部分组成的表达式可能在所有情况下都更快,因为两个比较可能是独立完成的,而在减法和比较方法中,减法必须在比较执行之前完成。
优雅是因为它不需要确定两个边界值中哪个先大。它也不包含分支。
public static bool InRange(float val, float a, float b)
{
// Determine if val lies between a and b without first asking which is larger (a or b)
return ( a <= val & val < b ) | ( b <= val & val < a );
}
当检查一个“数字”是否在一个范围内时,你必须清楚你的意思,两个数字相等意味着什么?一般来说,你应该把所有浮点数包装在一个所谓的“epsilon球”中,这是通过选择一个小的值来完成的,如果两个值如此接近,它们就是相同的。
private double _epsilon = 10E-9;
/// <summary>
/// Checks if the distance between two doubles is within an epsilon.
/// In general this should be used for determining equality between doubles.
/// </summary>
/// <param name="x0">The orgin of intrest</param>
/// <param name="x"> The point of intrest</param>
/// <param name="epsilon">The minimum distance between the points</param>
/// <returns>Returns true iff x in (x0-epsilon, x0+epsilon)</returns>
public static bool IsInNeghborhood(double x0, double x, double epsilon) => Abs(x0 - x) < epsilon;
public static bool AreEqual(double v0, double v1) => IsInNeghborhood(v0, v1, _epsilon);
有了这两个辅助,并假设任何数字都可以转换为double而不需要所需的精度。现在需要的是一个枚举和另一个方法
public enum BoundType
{
Open,
Closed,
OpenClosed,
ClosedOpen
}
另一种方法如下:
public static bool InRange(double value, double upperBound, double lowerBound, BoundType bound = BoundType.Open)
{
bool inside = value < upperBound && value > lowerBound;
switch (bound)
{
case BoundType.Open:
return inside;
case BoundType.Closed:
return inside || AreEqual(value, upperBound) || AreEqual(value, lowerBound);
case BoundType.OpenClosed:
return inside || AreEqual(value, upperBound);
case BoundType.ClosedOpen:
return inside || AreEqual(value, lowerBound);
default:
throw new System.NotImplementedException("You forgot to do something");
}
}
现在,这可能远远超过了您想要的,但它使您不必一直处理舍入问题,并试图记住一个值是否被舍入到哪个位置。如果你需要,你可以很容易地将它扩展到任意的情况并允许变化。
In production code I would simply write 1 <= x && x <= 100 This is easy to understand and very readable. Starting with C#9.0 we can write x is >= 1 and <= 100 Note that we must write x only once. is introduces a pattern matching expression where and is part of the pattern. && would require us to repeat x is as in x is >= 1 && x is <= 100 Here is a clever method that reduces the number of comparisons from two to one by using some math. There is not necessarily a performance advantage in doing so, but it is elegant. The idea is that one of the two factors becomes negative if the number lies outside of the range and zero if the number is equal to one of the bounds: If the bounds are inclusive: (x - 1) * (100 - x) >= 0 or (x - min) * (max - x) >= 0 If the bounds are exclusive: (x - 1) * (100 - x) > 0 or (x - min) * (max - x) > 0