How to use cellfun in only cells with numbers or how to create a csv file with different elements (numbers and strings)?

I have a big cell array (22002x12) with multiple elements (strings and numbers) and I want to convert them to a CSV file but obviously I can't. So I was using cellfun to convert everything in either strings or numbers but is not possible.

So my question is if I have a cell array like this:

 cellArray ={[        1]    '0.431089'      '0.210426'      '0.530389'      '0.529967'      '0.990659';  
             [        2]    '3.63584'       '2.9894'        '2.52868'       '3.80065'       '3.14567';     
             [        3]    [         0]    [         0]    '0.0593239'     [         0]    '0.110771';    
             [        9]    '1.51064'       '1.34827'       '1.77354'       '1.43471'       '1.21254';
             [       10]    '0.080619'      [         0]    '0.0531488'     '0.174925'      '0.0496248';  
             [       12]    '0.129531'      '0.169362'      '0.170763'      '0.21077'       '0.0797323'; 
             [       13]    '0.128055'      '0.133946'      '0.337633'      '0.208367'      '0.118236'}

how could I convert everything into numbers or cells using cellfun (or a better tool)?

I've been using :

 matrix = cellfun(@str2num, cellArray)

Or how can I create a CSV file with this data?

I've been using :

 csvwrite('new.csv',cellArray)

Thanks in advance!

Answer 1

[Stephen23]

This is easy with num2str, which ignores strings. This has the advantage that the output CSV file contains exactly the data in those strings, because there is no loss/rounding/... from converting to numeric and back again.

% Mixed cell array:
C ={[        1]    '0.431089'      '0.210426'      '0.530389'      '0.529967'      '0.990659';  
    [        2]    '3.63584'       '2.9894'        '2.52868'       '3.80065'       '3.14567';     
    [        3]    [         0]    [         0]    '0.0593239'     [         0]    '0.110771';    
    [        9]    '1.51064'       '1.34827'       '1.77354'       '1.43471'       '1.21254';
    [       10]    '0.080619'      [         0]    '0.0531488'     '0.174925'      '0.0496248';  
    [       12]    '0.129531'      '0.169362'      '0.170763'      '0.21077'       '0.0797323'; 
    [       13]    '0.128055'      '0.133946'      '0.337633'      '0.208367'      '0.118236'};
% Cell array of character vectors:
D = cellfun(@num2str,C.','uni',0);
% Save data in CSV file:
fmt = repmat(',%s',1,size(C,2));
fmt = sprintf('%s\n',fmt(2:end));
fid = fopen('temp0.csv','wt');
fprintf(fid,fmt,D{:});
fclose(fid);

and this creates the following CSV file:

1,0.431089,0.210426,0.530389,0.529967,0.990659
2,3.63584,2.9894,2.52868,3.80065,3.14567
3,0,0,0.0593239,0,0.110771
9,1.51064,1.34827,1.77354,1.43471,1.21254
10,0.080619,0,0.0531488,0.174925,0.0496248
12,0.129531,0.169362,0.170763,0.21077,0.0797323
13,0.128055,0.133946,0.337633,0.208367,0.118236

[German Preciat Gonzalez]

Thanks! exactly what I needed

Answer 2

[Andrei Bobrov]

l0 = cellfun(@ischar,cellArray);
cellArray(l0) = num2cell(str2double(cellArray(l0)));

or

l0 = cellfun(@isnumeric,cellArray);
cellArray(l0) = sprintfc('%d',[cellArray{l0}]);

Answer 3

[Guillaume]

You would be better off fixing whichever code you've used that creates a mixture of numbers and strings in the same column. You would also be better off using tables instead of cell array.

Failing that:

function x = maybeconvert(x)
   %convert x to number if it is a string, or leave it as is if already number
   if ischar(x)
     x = str2double(x);  %don't use str2num, it's a very dangerous function
end

You can use that with cellfun easily:

m = cellfun(@maybeconvert, cellArray);