How can i load .mat files from subfolders with for loop?

I have a different number of .mat files in several folders, which I would like to import, plot data und save it in the same folder as a pdf file.

Here an example with two folders:

C:\Users\I2d1_4_1_1\

y3.mat

y4.mat

y5.mat

C:\Users\I5d1_4_1_1\

y3.mat

y4.mat

y5.mat

y6.mat

My two questions are:

1. How do i save it in the same folder without changing the directory.
2. How to adapt the code that i dont need to change the for loop i=3:5 for the first folder and i=3:6 for the second folder.

DataPath ='C:\Users\';
for i=3:60 
for input=[2 5 8]
for delay=[4 8 12];
for delayinput=[1 4 8];               
    load([DataPath 'I%dd1_%d_1_%d\y' num2str(i) '.mat',input,delay,delayinput])   
figure(1), plotregression(x, y); 
SaveName = sprintf('NN_%d_1_%d',i,delay);
saveas(figure(1),fullfile(DataPath 'I%dd1_%d_1_%d', SaveName),'pdf')
end
end
end
end

Thanks for your help.

Answer 1

[Walter Roberson]

projectdir = 'C:\Users';
foldinfo = dir( fullfile(projectdir, 'I*') );
foldinfo(~[foldinfo.isdir]) = [];   %get rid of I* that are not folders
foldnames = fullfile(projectdir, {foldinfo.name});
numfold = length(foldnames);
for didx = 1 : numfold
    thisfold = foldnames{didx};
    dinfo = dir(fullfile(thisfold, 'y*.mat'));
    filenames = fullfile(thisfold, {dinfo.name});
    numfiles = length(filenames);
    for fidx = 1 : numfiles
        thisfile = filenames{fidx};
        datastruct = load(thisfile);
        x = datastruct.x; y = datastruct.y;
        and do your thing
    end
end

[Josef Bubendorfer]

Thanks, thats exactly what I was looking for.

I would now have two more questions.

How can I write the maximum of the data points x and y (sqrt (x2 + y2)) from each folder into a structure and read them into an Excel spreadsheet?

And the second question is how do I get all the values ​​in a plot (x, y) in which the folders are distinguished in the legend?

[Walter Roberson]

projectdir = 'C:\Users';
foldinfo = dir( fullfile(projectdir, 'I*') );
foldinfo(~[foldinfo.isdir]) = [];   %get rid of I* that are not folders
foldnames = fullfile(projectdir, {foldinfo.name});
numfold = length(foldnames);
all_min_x = zeros(1,numfold);
all_min_y = zeros(1,numfold);
for didx = 1 : numfold
    thisfold = foldnames{didx};
    [~, fold_name_only] = basename(thisfold);
    dinfo = dir(fullfile(thisfold, 'y*.mat'));
    filenames = fullfile(thisfold, {dinfo.name});
    numfiles = length(filenames);
    fold_min_x = inf; fold_min_y = inf;
    for fidx = 1 : numfiles
        thisfile = filenames{fidx};
        datastruct = load(thisfile);
        x = datastruct.x; y = datastruct.y;
        and do your thing
        fold_min_x = min(fold_min_x, x);
        fold_min_y = min(fold_min_y, y);
        plot(x, y, 'DisplayName', fold_name_only);
        hold on
        drawnow ratelimit
    end
    all_min_x(didx) = fold_min_x;
    all_min_y(didx) = fold_min_y;
end
now write all_min_x and all_min_y to your excel spreadsheet

[Josef Bubendorfer]

Thank you. But the following line is producing an error.

[~, fold_name_only] = basename(thisfold);

Undefined function 'basename' for input arguments of type 'char'.

I am using Matlab 2014a. Maybe because of that?

[Walter Roberson]

[~, fold_name_only] = fileparts(thisfold);

It was a mistake on my part. Too many years of working with too many languages.

[Josef Bubendorfer]

Now that is working but the line drawnow ratelimit gives an error as unknown command option.

I am really thankfull for your work.

Answer 2

[Akira Agata]

How about using dir function?

Using the function, the solution would be like this:

fileList = dir('C:\Users\**\*.mat');

for kk = 1:numel(fileList)
  filePath = fullfile(fileList(kk).folder,fileList(kk).name);
  load(filePath);
  h = figure;
  plotregression(x,y); 
  saveName = replace(fileList(kk).name,'.mat','.pdf');
  saveas(h,fullfile(fileList(kk).folder,saveName));
  close(h);
end

[Josef Bubendorfer]

Thanks for that answer. I think that would work if i would not have other files in the folders too such as other .mat files and .pdf files.

Sry for my inaccurate example. I wanted to keep it as simple as possible.

[Akira Agata]

OK, then you can skip unwanted .mat files by simply adding if statement.

For example, if you want to save in .pdf for only y1~y6.mat, the program would be like this:

fileList = dir('C:\Users\**\*.mat');
for kk = 1:numel(fileList)
    if ~isempty(regexp(fileList(kk).name,'y[1-6].mat'))
      filePath = fullfile(fileList(kk).folder,fileList(kk).name);
      load(filePath);
      h = figure;
      plotregression(x,y); 
      saveName = replace(fileList(kk).name,'.mat','.pdf');
      saveas(h,fullfile(fileList(kk).folder,saveName));
      close(h);
    end
end

[Image Analyst]

Or, for any y*.mat but not other files beginning with a different letter:

fileList = dir('C:\Users\**\y*.mat');

And then skip the "if" inside the "for".

You might also want to take a look at fileDatastore().

[Josef Bubendorfer]

Thanks for your help. I am usind Matlab 2014a and there the dir funktion this way is not working I think. Sry for the inaccurate question again.

[Akira Agata]

Then, you should change the first line to:

fileList = dir('C:\Users\I2d1_4_1_1\*.mat');

and change the fileList(kk).folder to 'C:\Users\I2d1_4_1_1\'.

Then, please do the same thing to the other folder C:\Users\I5d1_4_1_1\.

[Josef Bubendorfer]

OK, thanks. Thats what i more or less already do, but if there is no "automatical" way to do it I have to go on like this. But thanks for the help and your ideas.