Values in cell array keep getting overwritten
Show older comments
Hi there, I am new at MATLAB and currently trying to extract data from a particular subfolder and put it into a new cell array
basically i want to retrieve mean reaction times which is stored in a variable called a (location: parent_directory > containing many subject folders labelled img0001, img0002 etc > each subject folder contains a .mat file called stopchoicert.mat > variable a can be found here)
My code is as follows but instead of every iteration being placed into cell array, the values keep getting overwritten so i only have the mean reaction time of my last subject.
I have around 325 subjects but not all of them have mean reaction times, so I am unsure how many values in my cell array I should be getting at the end but the output rt that i get is a 326x1 cell array although only the 80th row contains the mean reaction time (of the last subject) and the rest of the cells are a 0X0 double
parent_directory = dir; %setting up the parent file to access files
rt = cell(size(parent_directory)); %an empty cell to insert all reaction times
for i = 1:length(parent_directory)
if ~isempty(strfind(parent_directory(i).name,'img')) %find folders with img in it
foldername = [parent_directory(i).folder '/' parent_directory(i).name]; %create a new path
cd(foldername); %loop through the paths
if isempty(dir('stopchoicert.mat'))
continue; %if folder does not contain stopchoicert.mat then skip
end
load ('stopchoicert.mat')
rt{i,1}= a %a contains the mean reaction time which i want to store in cell array
end %loop through all folders to extract all mean reaction times and store into cell array
end
rt
please help and lmk if you need any more info
Accepted Answer
Much simpler and much more robust:
D = 'path to the main directory';
S = dir(fullfile(D,'img*'));
for k = 1:numel(S)
F = fullfile(D,S(k).name,'stopchoicert.mat');
T = load(F);
S(k).data = T.a;
end
C = {S.data};
If not all of the subfolders contain the file, then you could do something like this:
for k = 1:numel(S)
F = fullfile(D,S(k).name,'stopchoicert.mat');
if exist(F,'file')==2
T = load(F);
S(k).data = T.a;
end
end
18 Comments
zexi Leong
on 18 Oct 2019
Edited: zexi Leong
on 18 Oct 2019
zexi Leong
on 18 Oct 2019
"could you explain to me what line 3 in your code means? why would it need to be ==7? what does the second input file mean? "
According to its documentation, the exist function will check for the "existence of variable, script, function, folder, or class". By default it searches for all of these, but we can help it by specifying the second optional argument: in this case 'file' tells exist to only check for files. The output from exist also depends on what it finds (i.e. variable, script, etc), and when you read the documentation you will find that an output value 7 corresponds to a file.
"This is my edited code based on your suggestion but i will get the error of Reference to non-existent field 'data'"
Your apparently random changes to my code are the reason why the data is not being imported.
- for some unknown reason you decided to add another exist call after loading the file data. Consider the two possibilities a) if the file does not exist, then trying to import its data will throw an error, so your check does nothing, or b) if the file does exist, then load has already imported its data, in which case your exist is total superfluous. That is why my code already checked if the file exists before loading its data.
- You included the part of the code that actually stores the imported data in an if statement that runs only when the file does not exist (how do you imagine storing data from non-existent files?)... but this is all nested inside my original if statement which already checks if the file exists: can you tell me how it is ever possible for both if statements to have true conditions? i.e. for your if statement to run, a file has to both exist and not exist. Perhaps you are checking for Schroedinger's file?
As far as I can tell, you made these random illogical changes because you were under the incorrect impression that the code needed a continue statement. If it needed a continue statement, I would have used one. I recommend actually looking at my code and thinking about its logic, especially how the if statement that I used simply runs if the file exists, and skips to the next loop iteration if the file does not exist (no continue is required).
" I am not too sure how to create the field 'data'?"
By using my answer, and not adding random changes that prevent it from working properly.
Stephen23
on 18 Oct 2019
zexi Leong's "Answer" moved here:
Hi sir, I am really grateful for the quick reply.
D = '/Users/jessieleong/Documents/MATLAB/crtgeoff_analysis copy 2';
S = dir(fullfile(D,'img*'));
for k = 1:numel(S)
F = fullfile(D,S(k).name,'stopchoicert.mat');
if exist(F,'file')==7
T = load(F);
S(k).data = T.a;
end
end
C = {S.data};
C
But even before my incorrect addition of the continue statement, i occured the same error.
Is this the correct path?
D = '/Users/jessieleong/Documents/MATLAB/crtgeoff_analysis copy 2';
You could just use D = '.'; to specify the current directory, but only if those subfolders are in the current directory (which appears to be the case, judging by your screenshot).
Note that the main advantage of using an absolute path is that your code and your data can be stored separately (which they should be).
zexi Leong
on 18 Oct 2019
Edited: zexi Leong
on 18 Oct 2019
@zexi Leong: I am not using a mac, so I cannot help with the correct syntax. You will have to check the path syntax yourself.
I am happy to help you to debug the code. Please show the output of these commands:
size(S)
exist(F,'file')
zexi Leong
on 18 Oct 2019
Edited: zexi Leong
on 18 Oct 2019
Note that S is empty, this means that dir did not find anything on the requested path that matched the requested name. This is often caused by using a wrong path (e.g. a spelling mistake), or looking in the wrong folder (e.g. one level higher than required).
So we need to find out why dir cannot find those subfolders.
Please show the output of these commands:
D = '/Users/jessieleong/Documents/MATLAB/crtgeoff_analysis copy 2';
dir(fullfile(D,'*'))
dir(fullfile('.','*'))
dir('*')
zexi Leong
on 18 Oct 2019
zexi Leong
on 18 Oct 2019
Stephen23
on 18 Oct 2019
" the '*' that you put in is used to look at any file in that directory right?"
Any file or subfolder.
Please try this:
D = '/Users/jessieleong/Documents/MATLAB/crtgeoff_analysis copy 2';
S = dir(fullfile(D,'*'));
size(S)
zexi Leong
on 18 Oct 2019
Stephen23
on 18 Oct 2019
@zexi Leong: why do you get different results when you run exactly the same code in your script vs. in the command window? Please run these commands in two locations:
- at the end of your script,
- in the command window.
And show the outputs:
pwd()
numel(dir(fullfile(D,'*')))
numel(dir(fullfile(D,'img*')))
zexi Leong
on 3 Dec 2019
Edited: zexi Leong
on 3 Dec 2019
Remember to accept my answer if it helped you: accepting answers is an easy way to show your appreciation of the volunteers who help you on this forum.
zexi Leong
on 3 Dec 2019
Stephen23
on 3 Dec 2019
"So i changed it from 7 to 2. This resolved all of my problems. "
Ah, well spotted! I will change my answer too, so that it does not confuse any future readers.
More Answers (1)
Divya Yerraguntla
on 16 Oct 2019
Hi Zexi
Try using the following modified code to serve your purpose. Have a look at the comments for a better understanding regarding the changes made.
parent_directory = dir;
num_dir = numel(parent_directory([parent_directory(:).isdir]))-2;% To find the number of sub folders on the parent directory
rt = cell(num_dir,1);
for i = 3:length(parent_directory) % The parent_directory.name field contains names of folders from the 3rd index location, hence i starts from 3
if ~isempty(strfind(parent_directory(i).name,'img'))
foldername = [parent_directory(i).folder '/' parent_directory(i).name];
cd(foldername);
if isempty(dir('stopchoicert.mat'))
continue;
end
load ('stopchoicert.mat');
rt{i-2,1}= a; % As i starts from 3, go back by 2 indices to fill rt
end
end
rt
Hope it helps!
1 Comment
This is factually incorrect: "The parent_directory.name field contains names of folders from the 3rd index location, hence i starts from 3"
As anyone experienced with MATLAB (or anyone who reads this forum) will know, the order of the files returned by dir is determined by the OS, and not by MATLAB. None of the OS MATLAB runs on guarantee a particular order (although apparently this author assumed that there is one).
The author incorrectly assumed that the first two results correspond to the folder shortcuts '.' and '..', but in fact it is trivial to define some filenames where this is not the case:
>> fclose(fopen('+test.txt','wt'));
>> fclose(fopen('-test.txt','wt'));
>> fclose(fopen('@test.txt','wt'));
>> S = dir('*');
>> S.name
ans = +test.txt
ans = .test.txt
ans = .
ans = ..
ans = @test.txt
Bad code includes
- the use of cd (slow, difficult to debug), rather than absolute/relative filenames.
- string concatenation instead of fullfile.
- loading directly into the workspace.
Categories
Find more on Shifting and Sorting Matrices in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!
