EDF writing - error when using cell array as signal DATA

Hello

I would like to do a edf file entering my signals from a cell array. I am doing a practice , mydata is

when I try to do this

>> C = {[1;2;3],[1;2;3;4],[1;2];

[4;5;6],[5;6;7;8],[3;4];

[7;8;9],[9;10;11;12],[5;6];

[10;11;12],[13;14;15;16],[7;8]}

>> hdr = edfheader("EDF");

hdr.NumSignals = 3;

hdr.NumDataRecords = 4;

hdr.DataRecordDuration = duration(0, 0, 1);

hdr.PhysicalMin = [1,1,1];

hdr.PhysicalMax = [12,16,8];

>> edfw = edfwrite('rand.edf',hdr,C);

However, I do get an error

Error using signal.internal.edf.write.validateSignalData

Expected input to be a vector

Error signal.internal.edf.write.getFileCreateOpts>checkSignals (第 223 行)

signal.internal.edf.write.validateSignalData(sigData, ...

Error signal.internal.edf.write.getFileCreateOpts>createHdr (第 96 行)

[reqhdr, sigData] = checkSignals(reqhdr, sigData, tNumSignals);

Error signal.internal.edf.write.getFileCreateOpts>createAndValidateHeader (第 47 行)

[hdr, sigData, tsal] = createHdr(thdr, sigData, tsal, hdrPropertiesList, ...

Error signal.internal.edf.write.getFileCreateOpts (第 13 行)

[hdr, sigData, tsal] = createAndValidateHeader(hdr, sigData, tsal, ...

Error edfwrite/createFile (第 1620 行)

signal.internal.edf.write.getFileCreateOpts(hdr, signalData, ...

Error edfwrite (第 499 行)

[obj, fileInfo] = createFile(obj, filename, hdr,...

Each column of my data represents a signal, and each signal is sampled at a different rate,

because I don't have specific data yet, so I did this exercise.

Answer 1

[Sandeep Mishra]

Hi Wensor,

I reviewed the code snippet and encountered a similar issue in MATLAB R2023b.

Upon further investigation, I discovered a relevant MATLAB Answers post which addresses a similar error: 

https://www.mathworks.com/matlabcentral/answers/1921840-edf-writing-error-when-using-cell-array-as-signal-data#answer_1186520

The root cause of this error is related to the input cell array 'C'. When using the ‘edfwrite’ function, if you specify a cell array as your signal data, each cell must correspond to a single signal.

However, the provided code snippet shows 'C' as a 4x3 cell array, with each cell containing vectors of varying sizes.

To resolve this issue, you can reshape the input signal 'C' to align with the expected format.

Here's a suggested approach:

% Reshaping input cell array C to C_reshaped
C_reshaped = cell(1,3); 
for col = 1:3 
    C_reshaped{col} = vertcat(C{:,col}); 
end 

% Using edfwrite with C_reshaped sigData
edfw = edfwrite('rand.edf',hdr,C_reshaped);

In the above approach, 'C_reshaped' will be a 1x3 cell array, where each column contains the concatenated results of all rows for that column, ensuring the data is structured correctly for ‘edfwrite’.

Please refer to the below documentation to learn more about ‘edfwrite’ function in MATLAB: https://www.mathworks.com/help/releases/R2023b/signal/ref/edfwrite.html

I hope this helps!

[Wensor]

Hi@Sandeep Mishra

That work perfectly! Thank you very much!