How to save data in cells to csv in loop

I have a struct with different fields from which I extract 120 columns of data in a loop. I prepend a number to each column and I'm now trying to save these columns to a csv file. All column have different lengths and I need the csv to be have these columns side-by-side. Also, I'm trying to use the row offset to prepend 6 empty rows to each column while writing to the csv. I tried to play with the parameters but csvwrite doesn't write the columns side-by-side. I'm not sure how to proceed from here.

n_chans = size(file.dat, 1); % get number of channels. Should be 120
for chan = 1:n_chans
    % get all timestamps from each channel
    timestamps = file.dat{chan, 1}.timestamps * 1000000; % convert seconds to microseconds
    timestamps = num2cell(timestamps);
    
    % get channel info
    full_name = file.dat{chan, 1}.name;
    underscore_indices = strfind(full_name, '_');
    channel_number = str2double(full_name(underscore_indices(3)+1 : underscore_indices(4)-1));
    
    % channel data as column
    column = [channel_number; timestamps]; % add the channel name/number
    
    % save data 
    csvwrite('test.csv', column, 6, col_offset);      % write with 6 row offset
    col_offset = col_offset + 1;
    
end

Answer 1

[Voss]

n_chans = size(file.dat, 1); % get number of channels. Should be 120

data = NaN(0,n_chans); % initialize matrix to be written to file
n_rows = 0; % number of rows in data, initially 0, will be increased below as necessary

column_offset = 1;

for chan = 1:n_chans

    % get all timestamps from each channel
    timestamps = file.dat{chan, 1}.timestamps * 1000000; % convert seconds to microseconds

    % get channel info
    full_name = file.dat{chan, 1}.name;
    underscore_indices = strfind(full_name, '_');
    channel_number = str2double(full_name(underscore_indices(3)+1 : underscore_indices(4)-1));

    % channel data as column
    column = [channel_number; timestamps]; % add the channel name/number

    % increase the number of rows of data as necessary
    nr = numel(column);
    if n_rows < nr
        data(n_rows+1:nr,:) = NaN;
        n_rows = nr;
    end

    % store column in data
    data(1:nr,col_offset) = column;

    col_offset = col_offset + 1;

end

% prepend 6 blank rows to data
data = [NaN(6,n_chans); data];

% write data to file
csvwrite('test.csv',data);
% or use writematrix
% writematrix(data,'test.csv');

[Francisco]

Thank you, I'm going to try it out. From what I understand the empty cells are populated with nan. Would there be a way to keep them empty?

[Francisco]

I'm getting the follwing error

Subscripted assignment dimension mismatch.

[Voss]

Sorry, I had a mistake. Please try the corrected code.

[Voss]

One way to have blank cells is to convert the matrix to a cell array, replace NaN element with [] (empty numeric array) or '' (empty character vector), and use writecell to write it

nandata = isnan(data);

data = num2cell(data);

data(nandata) = {[]};

writecell(data,'test.csv');

[Francisco]

Thank you again. I'll give it a go!

Answer 2

[Sulaymon Eshkabilov]

Why to use writetable()

[Francisco]

Sorry I did not understand. From what I gather writetable needs columns of the same size.

[Image Analyst]

If you have any more questions, then attach your data and code to read it in with the paperclip icon after you read this:

TUTORIAL: How to ask a question (on Answers) and get a fast answer