Matlab: str2double works on Windows PC but not on Apple Mac - matlab

Question

Say I have the following example file names: file_0250.pdf, file_0251.pdf, file_0252.pdf. I would like to get the following cell array:
'250 251 252'.
On the Windows PC at work I can run the following code with no problems, but on my Macbook at home I can not get the 'str2double' values as it returns a NaN value. It's frustrating:
folder_name = '/User/....';
file_name = 'file_';
extension = '.pdf';
%//' files pattern with absolute paths
filePattern = fullfile(folder_name, [file_name '*' extension] );
old_filename = cellstr(ls(filePattern)) ;
%// Get numbers associated with each file
file_ID = strrep(strrep(old_filename, file_name ,''), extension,'');
file_ID_doublearr = str2double(file_ID);
I tried 'cell2mat', 'str2mat', but they do not go well with the rest of the code:
file_ID_doublearr = file_ID_doublearr - min(file_ID_doublearr)+ start_number;
file_ID = strtrim(cellstr(num2str(file_ID_doublearr)));
%// Get zeros string to be pre-appended to each new_name
str_zeros = arrayfun(#(t) repmat('0',1,t), 4-cellfun(#numel,file_ID),'uni',0) ;
%// Generate new filenames
new_name = get(handles.new_name, 'string');
new_extension = get(handles.new_extension, 'string');
new_filename = strcat(new_name,str_zeros,file_ID,new_extension) ;
%// Finally rename files with the absolute paths
cellfun(#(m1,m2) movefile(m1,m2),fullfile(folder_name,file_name),fullfile(folder_name,new_filename)) ;

Answer 1

Your issue has to do with the different way that *nix (Linux and Mac) and Windows treat ls as mentioned in the documentation. As you've found out, ls returns a 2D character array of filenames. On the PC, these filenames will be returned one per row.
file_001.pdf
file_002.pdf
file_003.pdf
file_004.pdf
When you call cellstr on the result, it will place each filename into it's own cell array element after which you can successfully extract the number portion and convert them to numbers.
On *nix-based systems though, ls will typically yield a multi-column output. For example:
file_001.pdf file_002.pdf file_003.pdf
file_004.pdf
When you call cellstr on this, you will get one cell array element per row, but as you can see the first row actually contains three filenames. Then once you extract the number portion you would get something like this:
'001 002 003'
'004'
When you try to convert to a number, you're trying to convert a string of numbers to a single number and you get a NaN.
str2double({'001 002 003'; '004'})
% NaN 4
The best way to fix this is to not use the OS-dependent ls and use dir instead which is guaranteed to have consistent behavior across operating systems.
files = dir(fullfile(folder_name, [filename, '*', extension]));
numbers = regexp({files.name}, '[0-9]*', 'match');
The other option is to make sure that file_ID does not contain any space-separated numbers.
file_IDs = {'001 002 003'; '004'};
% Break each element up into multiple elements if it contains spaces
file_IDs = cellfun(#(x)strsplit(x), file_IDs, 'UniformOutput', 0);
file_IDs = cat(2, file_IDs{:});
% Now convert to a number
str2double(file_IDs);
% 1 2 3 4