CODEWITHSUNDEEP
matlabmatrixcombinations
user1884989
user1884989

Reputation: 41

All combinations for matrix in matlab

I'm trying find all combinations of an n-by-n matrix without repetitions.

For example, I have a matrix like this:

A = [321 319 322; ...
     320 180 130; ...
     299 100 310];

I want the following result:

(321 180 310)
(321 130 100)
(319 320 310)
(319 139 299)
(322 320 100)
(322 180 299)

I have tried using ndgrid, but it takes the row or the column twice.

Upvotes: 4

Views: 3177

Answers (3)

Eitan T
Eitan T

Reputation: 32920

Here's a simpler (native) solution with perms and meshgrid:

N = size(A, 1);
X = perms(1:N);                    % # Permuations of column indices
Y = meshgrid(1:N, 1:factorial(N)); % # Row indices
idx = (X - 1) * N + Y;             % # Convert to linear indexing
C = A(idx)                         % # Extract combinations

The result is a matrix, each row containing a different combination of elements:

C =

   321   180   310
   319   320   310
   321   130   100
   319   130   299
   322   320   100
   322   180   299

This solution can also be shortened to:

C = A((perms(1:N) - 1) * N + meshgrid(1:N, 1:factorial(N)))

Upvotes: 2

mythealias
mythealias

Reputation: 350

You can use perms to permute the columns as follows:

% A is given m x n matrix
row = 1:size( A, 1 );
col = perms( 1:size( A, 2 ) );

B = zeros( size( col, 1 ), length( row )); % Allocate memory for storage

% Simple for-loop (this should be vectorized)
% for c = 1:size( B, 2 )
%     for r = 1:size( B, 1 )
%         B( r, c ) = A( row( c ), col( r, c ));
%     end
% end

% Simple for-loop (further vectorization possible)
r = 1:size( B, 1 );
for c = 1:size( B, 2 )
    B( r, c ) = A( row( c ), col( r, c ));
end

Upvotes: 0

bonCodigo
bonCodigo

Reputation: 14361

ALLCOMB is the key to your question

E.g. I am not front of a MATLAB machine so, I took a sample from web.

x = allcomb([1 3 5],[-3 8],[],[0 1]) ;
ans
1 -3 0
1 -3 1
1 8 0
...
5 -3 1
5 8 0
5 8 1

Upvotes: 0

Related Questions