How can I set the FormatSpec to read this numerical value?

Question

Francesco Saltari 2018-12-7

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/434434-how-can-i-set-the-formatspec-to-read-this-numerical-value

回答： Francesco Saltari 2018-12-11

采纳的回答： dpb

Hi all,

I have some some difficulties to read from file the following numerical value by using textscan

'-1.661-4'

The problem lies in the lack of the exponential character that makes Matlab read -1.661.

Is there a simple way to set a FormatSpec to solve the problem. I tried with 'ExpChar ','' but it didn't work.

PS: this numerical value is within a loop where just few values have this structure.

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

dpb 2018-12-7

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/434434-how-can-i-set-the-formatspec-to-read-this-numerical-value#answer_351057

在 MATLAB Online 中打开

ML doesn't have the facility built into any of the text conversion routines to handle the omitted exponent indicator; some Fortran compilers have that as an extension to allow such a format.

One way if the file is regular in there always being the two elements is to write a little processing function...

s='-1.661-4,3.141+0';                                            % more general input string example
v=reshape(cell2mat(textscan(s,'%f','delimiter',',')),2,[]).';    % parse fields to mantiss/exponent
v=v(:,1).*10.^v(:,2);                                            % convert to value
>> fprintf('%g %g\n',v)      % show what we got
-0.0001661 3.141
>> 

If there are missing elements in some entries, then a parsing of each element and interpretation on a field-by-field basis would be necessary or write a more intelligent regular expression substitution pattern to correctly insert the missing exponent indication letter where it is needed.