How do I sort filenames containing text and numbers in numerical order in MATLAB?

I have a series of .png files in a folder whose names are the following;

(A)Test_Nom_BAUD1.png
(A)Test_Nom_BAUD7.png
(A)Test_Nom_BAUD8.png
(A)Test_Nom_BAUD10.png
(A)Test_Nom_BAUD11.png

In reading the MATLAB documentation, I understand the dir function sorts strings in ASCII dictionary order. And since the files I'm using have no leading zeros, this is a problem. This is exactly what I'm seeing when I use the dir function:

A = dir('*.png');

(A)Test_Nom_BAUD1.png
(A)Test_Nom_BAUD10.png
(A)Test_Nom_BAUD11png
(A)Test_Nom_BAUD7.png
(A)Test_Nom_BAUD8.png

But what I need is this;

A = dir('*.png');

    (A)Test_Nom_BAUD1.png
    (A)Test_Nom_BAUD7.png
    (A)Test_Nom_BAUD8.png
    (A)Test_Nom_BAUD10.png
    (A)Test_Nom_BAUD11.png

I know this can be done if the filenames contain only numbers.

But is it possible to sort these filenames (in numerical order) using existing text manipulation and sorting routines in MATLAB? If so, how?

[Stephen23]

"If so, how?"

Simply download my FEX submission natsortfiles, which was written to solve that exact problem:

S = dir('*.png');
S = natsortfiles(S); % alphanumeric sort by filename

Answer 1

[Cedric]

If, for any reason, you cannot install this function, you can sort your filenames as follows:

 [~, reindex] = sort( str2double( regexp( {A.name}, '\d+', 'match', 'once' )))
 A = A(reindex) ;

Note that it assumes that all files have the same base name. If you have to deal with various base names, then you need the function mentioned in the other answers, or to work a little more on the sorting.

[Brad]

GUH!! Cedric, how could I forget about regexp? I do have to deal with various base names, but that was a minor issue. Thank you for taking a look at this (and reminding me to keep using regexp).

[Cedric]

My pleasure!

[Hamidreza Etebariab]

I do have various base names and they look like this: 'RSN1111-KOBE-0.1_Node_Floor_Dsp.out'

I want to sort based on the 2 numbers, first sort based on the first number(1111) and then sort based on second one(0.1).

With your code I only can sort based on the first number. Can you please help me through this issue? Thanks!

[Image Analyst]

You'll have to write a custom parsing routine to extract out the numbers you want into numerical vectors then sort them yourself. Shouldn't be hard but let us know if you can't figure it out.

[Stephen23]

"You'll have to write a custom parsing routine to extract out the numbers you want into numerical vectors then sort them yourself."

You don't have to do that at all: natsortfiles handles multiple number values already. Just provide an appropriate regular expression, and it will work just fine:

>> C = {...
    'RSN1112-KOBE-1.1_Node_Floor_Dsp.out',...
    'RSN1111-KOBE-0.2_Node_Floor_Dsp.out',...
    'RSN1112-KOBE-0.2_Node_Floor_Dsp.out',...
    'RSN1111-KOBE-0.1_Node_Floor_Dsp.out',...
    'RSN1111-KOBE-1.1_Node_Floor_Dsp.out',...
    'RSN1112-KOBE-0.1_Node_Floor_Dsp.out'};
>> natsortfiles(C,'\d+\.?\d*') % alphanumeric sort
ans =
    'RSN1111-KOBE-0.1_Node_Floor_Dsp.out'
    'RSN1111-KOBE-0.2_Node_Floor_Dsp.out'
    'RSN1111-KOBE-1.1_Node_Floor_Dsp.out'
    'RSN1112-KOBE-0.1_Node_Floor_Dsp.out'
    'RSN1112-KOBE-0.2_Node_Floor_Dsp.out'
    'RSN1112-KOBE-1.1_Node_Floor_Dsp.out'

Answer 2

[KSSV]

Try this: https://in.mathworks.com/matlabcentral/fileexchange/10959-sort-nat--natural-order-sort

[Brad]

Thanks, but I can't get to that WEB site from this workstation.

[Jan]

Really? But it is on the same server as the forum. Perhaps it works if you insert your country code:

https://mathworks.com/matlabcentral/fileexchange/10959-sort-nat--natural-order-sort (For the US site, or "https://de." for Germany etc.

Answer 3

[Image Analyst]

See https://blogs.mathworks.com/pick/2014/12/05/natural-order-sorting/