problem reading with format

Hi,

I have a text file starting as

      1 19  2 3 1 5 4 1 5 6 1 7216.9023.30 2.7023.10 2.8017.40 1.1017.40 0.14 0.3484781 0.2780707 0.2950000 34.6  4.5 33.0 28.3 26.6 23.4 22.3 22.2 23.1 21.0 17.1 15.0 12.5  8.7  5.4  0.0  2.6  0.0
      1 23  1 3 1 1 2 1 2 6 5 6720.9027.20 2.2027.20 2.3022.90 1.6022.70 0.15 0.5713450 0.4988143 0.5600000 41.8  3.2 38.2 31.0 27.9 27.2 25.7 24.6 24.8 23.1 21.8 17.9 15.6 11.8  6.9  0.0  2.8  0.0

When I try to read it with format

form=[ '%7.0f %3.0f %3.0f %2.0f %2.0f %2.0f %2.0f %2.0f %2.0f %2.0f' ...
       '%2.0f %3.0f %5.2f %5.2f %5.2f %5.2f %5.2f %5.2f %5.2f %5.2f' ...
       '%5.2f %10.7f %10.7f %10.7f %5.1f %5.1f %5.1f %5.1f %5.1f %5.1f' ...
       '%5.1f %5.1f %5.1f %5.1f %5.1f %5.1f %5.1f %5.1f %5.1f %5.1f ' ...
       '%5.1f %5.1f']
c=textscan(fid,form,863);

Matlab does not read correctly. The field 12 is really 3 characters but Matlab reads in that place 4 characters, so that the 13. number is 6.9 but it should be 16.9. When I put a wrong format for field 12

form=[ '%7.0f %3.0f %3.0f %2.0f %2.0f %2.0f %2.0f %2.0f %2.0f %2.0f' ...
       '%2.0f %2.0f %5.2f %5.2f %5.2f %5.2f %5.2f %5.2f %5.2f %5.2f' .

then Matlab read the file correctly.

What might be the problem?

[per isakson]

See How can I use textscan without delimiter?

Answer 1

[Jeremy Hughes]

The leading spaces are not being counted toward the width of the fields as you're expecting.

%n.pf - will read at most n characters, starting at the beginning of a field, which does not include the spaces, it will include only the first p characters after the decimal.

You might have better luck with fixedWidthImportOptions. I didn't try this since I don't have your exact file.

>> w = [7 3 3 2 2 2 2 2 2 2 2 3 5 5 5 5 5 5 5 5 5 10 10 10 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5];
>> opts = fixedWidthImportOptions('NumVariables',numel(w),'VariableWidths',w);
>> opts = setvartype(opts,'double');
% In R2019a
>> A = readmatrix(file,opts);
% If using an earlier release, 
>> T = readtable(file,opts);
>> A = T.Variables;

Edit: Oops, forgot NumVariables.

[per isakson]

I have no luck with T = readtable(file,opts); on R2018a neither with leading spaces in the text file nor without leading spaces

K>> getReport(ME)  % at parseInputs (line 28+2)
ans =
    'Error using matlab.io.text.FixedWidthImportOptions/set.VariableWidths (line 92)
     Expected a vector of positive integers of the same length as VariableNames.
     Error in matlab.io.internal.mixin.HasPropertiesAsNVPairs/parseInputs (line 28)
                         obj.(param{:}) = results.(param{:});
     Error in matlab.io.text.FixedWidthImportOptions (line 74)
                 [opts,otherArgs] = opts.parseInputs(varargin,{'NumVariables','VariableOptions','VariableNames'});
     Error in fixedWidthImportOptions (line 37)
     opts = matlab.io.text.FixedWidthImportOptions(varargin{:});
     Error in LiveEditorEvaluationHelperESectionEval (line 20)
     opts = fixedWidthImportOptions('VariableWidths',[1 3 3 2 2 2 2 2 2 2 2 3 5 5 5 5 5 5 5 5 5 10 10 10 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5])
     Error in matlab.internal.editor.evaluateCode'
     

More luck with the interactive Import Data

Isn't there a m-code generation option?

[Jeremy Hughes]

Wrote code without running it /shame.

I forgot to add NumVariables. Edited.

Import Tool generates MATLAB code for fixed width files, but I think there are limits to what it can do.