如何在Matlab矩阵的每一行中删除多余的重复元素？

Question

Let's say I have a matrix

A = [2 3 2 5 6 7 2; 
     1 2 5 4 5 6 7; 
     7 5 3 9 8 1 2];

How do I remove 2 s and keep one 2 in the first row and keep only one 5 in the second row?

Answer 1

The result can't be a matrix anymore, because each row will have a different length. You can obtain the result as a cell array of row vectors as follows:

B = mat2cell(A, ones(size(A,1),1)); %// convert matrix to cell array of its rows
B = cellfun(@(x) unique(x,'stable'), B, 'uniformoutput', 0); %// stably remove duplicates

For your example matrix

A = [2 3 2 5 6 7 2; 
     1 2 5 4 5 6 7; 
     7 5 3 9 8 1 2];

this gives

B{1} =
     2     3     5     6     7
B{2} =
     1     2     5     4     6     7
B{3} =
     7     5     3     9     8     1     2

Answer 2

If you want to find out which values are duplicates within the row, you can do something like this:

[vals, col_idx]  = sort(A,2);
idx = bsxfun(@plus,(col_idx-1)*size(A,1), (1:size(A,1))');
is_duplicate(idx(:,2:end)) = vals(:,1:end-1) == vals(:,2:end);
is_duplicate = reshape(is_duplicate, size(A));

is_duplicate =

 0     0     1     0     0     0     1
 0     0     0     0     1     0     0
 0     0     0     0     0     0     0

From there, it depends what outcome you are looking for. You could set the duplicates to NaN or some other value, or you could set them to NaN , but then shift them to the end of the row, using something like the following:

col_idx = cumsum(~is_duplicate, 2);
idx = bsxfun(@plus,(col_idx-1)*size(A,1), (1:size(A,1))');
A_new = nan(size(A));
A_new(idx(~is_duplicate)) = A(~is_duplicate);

A_new =

 2     3     5     6     7   NaN   NaN
 1     2     5     4     6     7   NaN
 7     5     3     9     8     1     2