CODEWITHSUNDEEP
matlabmatrix
EyesOfÖzil
EyesOfÖzil

Reputation: 309

How to get unique pairs of numbers in matlab where both numbers have not repeated in the matrix before

I have an input matrix as below

all = [12 16;12 13;8 14;14 19;3 6;8 6;13 25;25 14;7 2];

I need the following output

output = [12 16;8 14;3 6;13 25;7 2];

The explanation for the output is as follows.

First row of input i.e. 12 16 is the first row in output as both the numbers have never been repeated before in the output matrix (obviously).

Second row of input i.e 12 13 is not needed as the number 12 is present in first row of output i.e repeated

Third row of input i.e 8 14 is second row of output as both the numbers have never been repeated before in the output matrix.

Fourth row of input i.e 14 19 is not needed as the number 14 is present in output i.e repeated

On similar lines 3 6 needed as both are not repeated,

8 6 not needed as both 8 and 6 are repeated,

13 25 needed as both are not repeated

25 14 not needed as both are repeated

7 2 needed as both are not repeated

I am not able to get any ideas to start. Any help will be appreciated.

Thanks!

Upvotes: 2

Views: 152

Answers (2)

Luis Mendo
Luis Mendo

Reputation: 112659

This assumes that if a row contains two equal values they count as repeated and thus the row should be removed.

Don't use all as a variable name, because that shadows a function:

A = [12 16;12 13;8 14;14 19;3 6;8 6;13 25;25 14;7 2]; % input matrix
[~, u] = unique(A.', 'first'); % unique values of linearized transposed A.
    % In recent Matlab versions you an remove 'first '
M = false(flip(size(A))); % initiallize mask of values to be kept
M(u) = true; % fill values
output = A(all(M,1),:); % keep rows that only have non-repeated values

This gives

output =
    12    16
     8    14
     3     6
     7     2

Upvotes: 1

ibezito
ibezito

Reputation: 5822

One Liner Solution

res = all(arrayfun(@(ii) isempty(intersect(all(1:ii-1,:),all(ii,:))),1:size(all,1)),:);

Result

res =

12    16
 8    14
 3     6
 7     2

Explanation

let's divide the one-liner into a more detailed and documented chunk of code:

%defines a function which validates for each index wheter the row is
%completely unique are not.
uniqueRowIndicator = @(ii) isempty(intersect(all(1:ii-1,:),all(ii,:)));

%finds all the unique row in the matrix
inds = arrayfun(uniqueRowIndicator,1:size(all,1));

%extracts the result from the returned indices
res = all(inds,:);

Upvotes: 2

Related Questions