Extract value from every struct

I have four 1x1 struct files like the attached. mat file in every 60 subfolder, I want to extract the u field in each struct and assign its name to it. The name is in a field called name e.g, the name of this one is rdf. Then use it as one of the X e.g (model = fitcsvm(X,Y, 'KernelFunction','linear', 'BoxConstraint',1);) for svm classification. where, X= [X1 X2 X3 X4] and X1 is currently rdf. Using a for loop, please how do I extract something from a struct.

Answer 1

[Walter Roberson]

vois={'ldF', 'rdF', 'lvF', 'rvF'};
numvois = length(vois);

basedir = 'place/where/the/folders/are';
foldinfo = dir(basename);
foldinfo(~[foldinfo.isdir]) = [];  %remove non-folders
foldinfo(ismember({foldinfo.name}, {'.', '..'})) = [];  %remove . and ..
foldernames = fullfile({foldinfo.folder}, {foldinfo.name});
numfolders = length(foldernames);

X_all = cell(numfolders, numvois);

for K = 1 : numfolders
    thisfolder = foldernames{K};
    [~, basename, ~] = fileparts(thisfolder);
    for vidx = 1 : numvois
        filename = fullfile(thisfolder, basename + "_" + vidx + ".mat");
        Features = load(filename);
        X_i = Features.xY.u;
        X_all{thisfolder, vidx} = X_i;
    end
end

At the end of this, X_all will be a cell array with as many rows as folders (60, you said), and as many columns as vois (4)

[Monalisa Chikezie]

Hi, this gave me an error. I'm trying to find out why. however, each vois xY.u is 1 * 198 double, subfolder= 60 and there are 4 vois in total.

[Walter Roberson]

what error do you see.

[Monalisa Chikezie]

File: loading_data_test.m Line: 15 Column: 71

Invalid expression. When calling a function or indexing a variable, use parentheses. Otherwise, check for mismatched

delimiters.

[Walter Roberson]

Ah, I fixed it.

[Monalisa Chikezie]

oh, thanks a lot

Answer 2

[Karim]

You can acces the data in the struct by placing a point after the variable name of the struct, follewed by the fieldname you want to acces. See below for an example for the 'u' variable.

Features = load(websave('myFile', "https://nl.mathworks.com/matlabcentral/answers/uploaded_files/1085720/VOI_rdF_1.mat"));

% first extract xY from 'Features'
xY = Features.xY;
% then extract u from the struct
u = xY.u;

% plot the data
figure
plot(u)
grid on

Answer 3

[Monalisa Chikezie]

Hi, thanks for the respnse.

this is what I have so far, but it seems to give an error.

vois={'ldF', 'rdF', 'lvF', 'rvF'};
for j=1:4
    Features = load(pwd, '_1.mat'));
    X_i=Features.xY.u;
    X_all(:)=X_i;

[Karim]

FYI, you can just comment below the answer, you do not need to add an extra answer ;)

I did not know that you loaded the data into another variable, I adjusted my answer accordingly.

[Walter Roberson]

no definition of Y in this code.

[Monalisa Chikezie]

Y, here was an error. I've corrected it and edited the question for better explanation of what I am trying to do

[Walter Roberson]

When you assign to X_all(:) and X_all does not already exist then the resulting size is the same as the right hand side. This syntax for an initial assignment does not specifically create a column vector.

When you assign to X_all(:) and the right hand side does not have exactly the same number of elements as X_all has, then you will get an error.

If the two sides have the same number of elements exactly then the values would be copied in linear index order even if the two sides had different shapes.

I predict that the files contain different data sizes and your way of overwriting all of X_all in place is causing failure

[Monalisa Chikezie]

Alright, thank you. How do I extract it properly without errors

[Walter Roberson]

What was your intention in assigning to X_all(:) ?

[Monalisa Chikezie]

to capture all of the 4 vois in each of the 60 folders