C ++ Precision的问题

Question

我看到许多帖子谈到了如何在C ++中以给定的精度显示数字，但是我想知道是否有一种方法可以显示非重复数字，例如x.0和setprecision指定的宽度重复数字方法？

Answer 1

If you knew in advance whether the number you want to print exhibits a recurring decimal expansion, then you could simply issue an statement along the lines of:

if is_recurring:
    print using expanded format
else
    print using fixed format

I imagine you do not know in advance. And the computer cannot, in general, know it either.

How would it test whether a few decimals (about 15 if you are using double precision) represent a recurring decimal?

Furthermore, how would you like to print it? Take, for instance, the classical example:

3227 / 555 = 5.814414414414414

How should I print it? Like this?

5.8144

or like that?

5.8144
   ***

It seems like a very difficult problem to address in general.

Answer 2

It sounds like what you really want is to trim trailing zeroes. The machine doesn't know whether a number is really a "recurring number," which is just a fancy name for a rational number whose denominator is relatively prime from the numeric base, because floating point math is not based on rational numbers.

As Adam mentioned, the default output style is already to omit trailing zeroes. This hasn't much to do with math; it's just deleting the characters from the end of the string.

In terms of the Standard, the ios_base::fixed and ios_base::scientific flags determine qualitative formatting. If neither is set, the output is equivalent to the %g format specifier of printf .

According to man printf ,

Trailing zeros are removed from the fractional part of the result; a decimal point appears only if it is followed by at least one digit.

To get a .0 at the end despite the above rule, you would have to do something like print to a temporary string, search it for a decimal point, and retry in fixed mode if not found.