Rearranging filenames with extension .number (Ordering)

4 views (last 30 days)
I've got batches of data that end in MRDC.number (MRA slices e.g. i3078011.MRDC.205) and I want to order them by filename extensions i.e. MRDC.20,MRDC.21,MRDC.22,MRDC.23
Trouble is sort_nat for example reads the initial numbers not the extension
They go from MRDC.1 to MRDC.200ish ,but may vary
I've attached a typical filename cell array.
  1 Comment
William on 24 Apr 2015
This is a far as I've got
for i = 1:size(filen)
[~,name(i),ext(i)] = fileparts(char(filen(i)));

Sign in to comment.

Accepted Answer

Guillaume on 24 Apr 2015
Edited: Guillaume on 24 Apr 2015
Building up on Ingrid's answer and your code, this will sort the extensions numerically, regardless of the filename:
ext = regexp(filen, '(?<=\.)\d+$', 'match', 'once');
[~, order] = sort(str2double(ext));
filen = filen(order)
I use a regular expression instead of fileparts so I can get rid of the . as well.

More Answers (2)

Ingrid on 24 Apr 2015
if all your filenames have the same length (which seem to be the case for the file in attachment) you could do this
fin = cellfun(@(x) x(15:end),filen,'UniformOutput',false);
fin = str2double(fin);
[~,idx] = sort(fin);
sortedFileNames = filen(idx);
  1 Comment
Ingrid on 24 Apr 2015
alternativaly, if the size of the filename is not the same for all files you could use this
fin = cellfun(@(x) regexp(x,'\.','split'),filen,'UniformOutput',false);
fin = cellfun(@(x) x{end},fin,'UniformOutput',false);
fin = str2double(fin);
[~,idx] = sort(fin);
sortedFileNames = filen(idx);

Sign in to comment.

Stephen23 on 24 Apr 2015
Edited: Stephen23 on 24 Apr 2015
There are several natural order sorting functions on MATLAB File Exchange, and they often use file-names in their examples... but they all suffer from the same problem: that extension separator character just keeps getting in the way!
To avoid this problem I created natsortfiles, which sorts the file-names and file-extensions separately. Take a look at the code and you will find it easy to adapt for your purposes.
It never occurred to me that someone might want to sort only by file-extension, but that would be a useful option to add...
Alternatively, because the extensions consist entirely of digits, simply split these from the names and sort them numerically:
>> load('filenames.mat')
>> [pth,nam,ext] = cellfun(@fileparts,filen,'UniformOutput',false);
>> [~,idx] = sort(cellfun(@(s)sscanf(s,'.%i'),ext));
This uses the built-in fileparts function, so it is not depended on the file-name lengths or anything like that, and the sscanf call allows us to neatly ignore the period character. Two lines of code and you have your sort index!


Find more on File Name Construction 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!