Finding substring based on another substring

I have a bunch of data files and can get all of their names via

D = dir('*.mat');

Each name has a couple of numbers in it in the format '*num1_###*num2_###*.mat' where ### is the actual number.

I can set num1 equal to some value, say 1.7, and then get all assotiated files i.e.

D = dir('*num1_1.7*)

essentially getting a row of num2 files for num1 = 1.7 if you think of num1 and num2 making a grid of values.

What is the most efficient way to get files along the diagonal? I.e. where num1 is equal to num2, but the values aren't known a priori.

For example, if I have the files

num1_1_num2_1.mat num1_2_num2_1.mat num1_3_num2_1.mat

num1_1_num2_2.mat num1_2_num2_2.mat num1_3_num2_2.mat

num1_1_num2_2.mat num1_2_num2_3.mat num1_3_num2_3.mat

from dir(), how can I extract

num1_1_num2_1.mat, num1_2_num2_2.mat, num1_3_num2_3.mat?

Answer 1

[Stephen23]

"What is the most efficient way to get files along the diagonal?"

Probably something like this is about as efficient as you can get with MATLAB:

S = dir('*.mat');
N = {S.name};
M = sscanf([N{:}],'num1_%d_num2_%d.mat',[2,Inf])
Z = N(M(1,:)==M(2,:))