Show limited digits numbers.

Question

armin m 2021-10-15

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1564926-show-limited-digits-numbers

评论： armin m 2021-10-16

Hi i wanna numbers limited to 6 digits. For example. A=2.54357766 be show A=2.543578. Tnx

2 个评论
显示无隐藏无

Cris LaPierre 2021-10-16

Where do you want this to be shown?

armin m 2021-10-16

编辑：armin m 2021-10-16

Command windows and every where i wanna to print. E.g .txt file from diary on mode. But my main purpose relate to calculation. I have ref. Which calculated accordding to 6 digits numbers so i think the numbers with for example 10 digits have effect on result. So i wanna decrease digits to 6 to compare my results with reference.

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Chunru 2021-10-16

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1564926-show-limited-digits-numbers#answer_809746

You can use fprintf with format specifier to format the print out. "doc fprintf".

A=2.54357766;
fprintf('%.6f\n', A);
2.543578

2 个评论
显示无隐藏无

Walter Roberson 2021-10-16

Note that doing so will not change the result. If you are concerned that "10 digits have effect on results" and you want to decrease the digits to 6 for calculation purpose, then you are not going to achieve your aims just by using fprintf()

Perhaps what you are looking for is

format long g
A=2.54357766
A = 
                2.54357766
B = round(A, 6)
B = 
                  2.543578

If so, then be careful: MATLAB does not use decimal calculations, and 0.1 is not exactly reprepresentable in binary calculations, so round(,6) is not actually rounding to 6 decimal places, only to as close as it can get:

fprintf('%.999g\n', B)
2.543578000000000116642695502378046512603759765625

That is the exact binary representation that is used.

If you need calculations done in 6 decimal places, then you have three choices:

Use the Symbolic Toolbox... and worry about the effect of the 5 hidden guard digits during calculations; or
Calculate in integer ratios, perhaps using John D'Errico's Variable Precision Integer package; or
Use the Fixed Point Toolbox