How to convert nested structure to contain its original data?

I'm new, and I need some help. So I have these multiple signals in double, .mat format and I managed to load them all as output using this simple code. And I want to further analyze this dataset.

But the output sets the values as struct, which doesn't allow me to conduct batch processing such as feature extraction on it, or import it in the Diagnostic Toolbox.

I want the data in the cell array to be as double, not struct. May I know how to solve this? Or it is bc there's something with my data?

[Mathieu NOE]

hello Aisya

how are the data in the mat file organized ?

how would you like the mydata format to be ?

[Aisya Amelia]

hello Matthieu! i'll try my best to explain with pictures since i'm still a beginner. so each .mat file that I have is a GSR signal, where each have different number of samples. so each .mat file has these numeric values (variables) that i would like to extract features such as mean etc. my initial plan is load all these signals and output is as a variable so i can conduct batch processing instead of manually extracting features from each .mat file.

when i loaded the files, it stored each .mat as struct.

i want mydata to store the variables of each .mat file as it is (double), not struct.

[Mathieu NOE]

Well

I would think that the batch process could be done much simpler if you do the load and all computations in the same loop instead of generating another big database. My 2 cents !!

[Aisya Amelia]

thank you! but i'm still very much confused. if i do the load and all computations in the same loop, is there a way to output the variables of of each .mat file as it is or cell? im sorry if my words are confusing. lets say each participant, there is 16 signals. as example, this is dataset for participant 1 before i pre-process which i had downloaded online. i'm confused on how they were able to store it as such.

for this cell above, the matlab generated code are: i had removed some columns where there were redundant data.

function importfile1(fileToRead1)
%IMPORTFILE1(FILETOREAD1)
%  Imports data from the specified file
%  FILETOREAD1:  file to read

%  Auto-generated by MATLAB on 15-Apr-2021 08:48:22

% Import the file
newData1 = load('-mat', fileToRead1);

% Create new variables in the base workspace from those fields.
vars = fieldnames(newData1);
for i = 1:length(vars)
    assignin('base', vars{i}, newData1.(vars{i}));
end

to pre-process it, i had to extract out each signal individually. example for signal 1, i used:

P1 = GSR_DATA(1,1) and then i used the signal analyzer toolbox to perform pre-processing. and i exported every signal into a new .mat file and saved in a new folder. now, i would like to combine all these exported signals into one dataset as previously. i attempted to load all the dataset as shown earlier and it came out as struct.

my code:

myFolder1 = ('C:\Users\User\Documents\Signals');
 
 
mysignals=dir('**/*.mat');
N = length(mysignals);
mydata=cell(N,1);
 
for k = 1:N
    mydata{k}=load(mysignals(k).name);
end

thank you!

[Mathieu NOE]

hello

maybe you should start expalin first what you intend to do on processing the data , then I can help you to put together a code

maybe you could share some mat file so I can start work on this

tx

[Aisya Amelia]

Here’s a link to a set of signals. signals

I want to organize these signals into a single cell array or table (as long its not struct), smthg like this:

so that I could extract values such as mean, rms simultaneously instead of extracting one by one. Then I would like to store all these extracted features in another set of array/table. thank you!

[Stephan]

Did you try to use:

load(mysignals(k).name);

instead of

mydata{k}=load(mysignals(k).name);

This should give you the data in the way it was saved. if you assign the result of load to a variable, you will end up with a struct.

[Jan]

"in another set of array/table" - please do not let the reader decide, if you want a table or array. This is your decision.

[Aisya Amelia]

Hello @Stephan ! I did try after you suggested it, but it ended up loading all .mat files as separate variables in the workspace: when i opened mydata, it was just empty, [ ] . If I do not assign the result of load to a variable, how do I load them as one output? thank you!

Answer 1

[Jan]

myFolder1 = ('C:\Users\User\Documents\Signals');
mysignals = dir(fullfile(myFolder1, '**', '*.mat'));
N = numel(mysignals);
mydata = cell(N,1);
for k = 1:N
    File = fullfile(mysignals(k).folder, mysignals(k).name);
    Data = load(File);
    mydata{k} = Data.P2V1;  % Or P2V3 or what ever?
end

[Aisya Amelia]

hello! thank you for this code. i tried this, but up to this part as shown below, i got confused. It gave me an error Reference to non-existent field 'P2V1'. It is because of the naming of each of my signal files?

mydata{k} = Data.P2V1;  % I got confused here

This is how mydata came out to be, and workspace:

And also referring to your comment above "please do not let the reader decide, if you want a table or array. This is your decision." I truly apologize for not being firm. I would like it to be a cell.

[Jan]

Which data do you want to import from the MAT files? What is the contents of these files?

[Aisya Amelia]

Each .mat file is a signal of m x 1 cell array. The variable of each signal are of numeric values (conductance). I want to import these arrays of values.

[Jan]

All we need to knwo for the solution is the name of the variable stored in the MAT file. Replace my "P2V1" by the name of this field. I've written "or what ever" in the hope, that you can adjust it accordingly.