一般来说,数值型别的极限是一个与平台相关的特性。c++标准程序库通过template numeric_limits提供这些极值,取代c语言所采用的预处理常数。你仍然可以使用后者,其中整数常数定义于<climits>和<limits.h>,浮点常数定义于<cfloat>和<float.h> 。新的极值概念有两个优点,第一提供了更好的型别安全性,第二是程序员可借此写出一些template以核定这些极值。
1.内建型别的最小长度
C++Standard 规定了各种型别必须保证的最小精度:见下:
型别 最小长度
char 1 byte
short int 2bytes
int 2bytes
long int 4bytes
float 4bytes
double 8bytes
long double 8bytes
下面是C++11对numeric_limits的定义:(来自http://www.cplusplus.com/reference/limits/numeric_limits/)
1 template <class T> class numeric_limits { 2 public: 3 static constexpr bool is_specialized = false; 4 static constexpr T min() noexcept { return T(); } 5 static constexpr T max() noexcept { return T(); } 6 static constexpr T lowest() noexcept { return T(); } 7 static constexpr int digits = 0; 8 static constexpr int digits10 = 0; 9 static constexpr bool is_signed = false; 10 static constexpr bool is_integer = false; 11 static constexpr bool is_exact = false; 12 static constexpr int radix = 0; 13 static constexpr T epsilon() noexcept { return T(); } 14 static constexpr T round_error() noexcept { return T(); } 15 16 static constexpr int min_exponent = 0; 17 static constexpr int min_exponent10 = 0; 18 static constexpr int max_exponent = 0; 19 static constexpr int max_exponent10 = 0; 20 21 static constexpr bool has_infinity = false; 22 static constexpr bool has_quiet_NaN = false; 23 static constexpr bool has_signaling_NaN = false; 24 static constexpr float_denorm_style has_denorm = denorm_absent; 25 static constexpr bool has_denorm_loss = false; 26 static constexpr T infinity() noexcept { return T(); } 27 static constexpr T quiet_NaN() noexcept { return T(); } 28 static constexpr T signaling_NaN() noexcept { return T(); } 29 static constexpr T denorm_min() noexcept { return T(); } 30 31 static constexpr bool is_iec559 = false; 32 static constexpr bool is_bounded = false; 33 static constexpr bool is_modulo = false; 34 35 static constexpr bool traps = false; 36 static constexpr bool tinyness_before = false; 37 static constexpr float_round_style round_style = round_toward_zero; 38 };
numeric_limits<>的成员:
digits10 |
返回目标类型在十进制下可以表示的最大位数 |
epsilon |
返回目标数据类型能表示的最逼近1的正数和1的差的绝对值 |
has_denorm |
测试目标类型是不是可以非规范化表示示 |
has_denorm_loss |
测试所有类型是不是能测出因为非规范化而造成的精度损失(不是因为结果本身的不精确) |
has_infinity |
测试目标类型是不是能表示无限(比如被0除,或者其他一些情况) |
has_quiet_NaN |
检查目标类型是不是支持安静类型的NaN |
has_signaling_NaN |
检查目标类型是不是支持信号类型的NaN |
infinity |
检查目标类型的无限类型(如果支持无限表示) |
is_bounded |
检查目标类型的取值是否有限 |
is_exact |
测试目标类型的计算结果是不是不会造成舍入误差(比如float是0) |
is_iec559 |
测试目标类型是不是符合IEC559标准 |
is_integer |
测试目标类型是不是可以用整型来表示(比如char是1,float是0) |
is_modulo |
Tests if a type has a modulo representation. |
is_signed |
测试目标类型是否是带符号的 |
is_specialized |
测试目标类型是不是在numeric_limits .模板类中有特殊定义 |
max |
返回可取的有限最大值 |
max_exponent |
Returns the maximum positive integral exponent that the floating-point type can represent as a finite value when a base of radix is raised to that power. |
max_exponent10 |
Returns the maximum positive integral exponent that the floating-point type can represent as a finite value when a base of ten is raised to that power. |
min |
返回可取的最小值(规范化) |
min_exponent |
Returns the maximum negative integral exponent that the floating-point type can represent as a finite value when a base of radix is raised to that power. |
min_exponent10 |
Returns the maximum negative integral exponent that the floating-point type can represent as a finite value when a base of ten is raised to that power. |
quiet_NaN |
返回目标类型的安静NAN的表示 |
radix |
Returns the integral base, referred to as radix, used for the representation of a type. |
round_error |
返回目标类型的最大可能的舍入误差 |
round_style |
Returns a value that describes the various methods that an implementation can choose for rounding a floating-point value to an integer value. |
signaling_NaN |
返回目标类型关于信号NAN的表示 |
tinyness_before |
测试目标类型是不是能测定出微小的舍入误差 |
traps |
Tests whether trapping that reports on arithmetic exceptions is implemented for a type. |
例子:
#include<iostream>
#include <limits>
#include
<string>
using namespace std;
int main()
{
cout
<< boolalpha;
cout << "max(short): "<<
numeric_limits<short>::max() <<endl;
cout <<
"max(int): " << numeric_limits<int>::max() <<endl;
cout << "max(long): " << numeric_limits<long>::max() <<
endl;
cout <<endl;
cout << "max(float): "
<< numeric_limits<float>::max() <<endl;
cout
<< "max(double): "<< numeric_limits<double>::max()
<<endl;
cout << "max(long double): " <<
numeric_limits<long double>::max() <<endl;
cout <<
endl;
cout << "is_signed(char): " <<
numeric_limits<char>::is_signed <<endl;
cout <<
"is_specialized(string): "
<<numeric_limits<string>::is_specialized<<endl;
return 0;
}
在我台电脑上的运行结果为:
max(short): 32767
max(int): 2147483647
max(long):
2147483647
max(float): 3.40282e+38
max(double):
1.79769e+308
max(long double): 1.18973e+4932
is_signed(char):
true
is_specialized(string): false