Find back lost values from an original matrix of derived incomplete column vectors

I have a matrix M and I am extracting a submatrix A of selected column vectors from M , without keeping the values of the third row(substituted by nan). The selection method loses the 3rd row values of columns.

M=[1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2;1,1,1,1,2,2,2,2,4,4,4,1,1,1,2,2,2;1,2,3,4,1,2,3,4,1,2,3,2,3,4,1,2,3;31,16,15,9,10,8,13,6,16,11,4,9,4,10,6,17,15];

A= [1,1,2,2,2;1,2,4,1,2;NaN,NaN,NaN,NaN,NaN;31,13,16,10,17]

Is there a way to retrieve their values of their third row from M specifically?

The result would be R=

1 1 2 2 2

1 2 4 1 2

1 3 1 4 2

31 13 16 10 17

or

R=[1,1,2,2,2;1,2,4,1,2;1,3,1,4,2;31,13,16,10,17]

Thank you

Answer 1

[Awais Saeed]

You did not mention how are you picking the columns. I am assuming that you want columns 1 7 9 14 16 from M. If that's the case then

R = M(:,[1 7 9 14 16])

R =

1 1 2 2 2

1 2 4 1 2

1 3 1 4 2

31 13 16 10 17

[Awais Saeed]

This can also be done in a loop

v = [1 7 9 14 16]; % extract these columns from M
for col = 1:1:size(v,2)
    R(:,col) = M(:,v(col));
end
reshape(R', 1, [])

[Radu Mihail]

No Awais,

I want a method to detect which columns of M were selected in A submatrix, without you picking them up by inspecting M. Basically having M and A the method needs to find R.

What you have done can not be done if you have a huge M matrix with 1000 columns. My M here was truncated to illustrate what the method needs to be done

[Awais Saeed]

This will do the work

M=[1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2;1,1,1,1,2,2,2,2,4,4,4,1,1,1,2,2,2;1,2,3,4,1,2,3,4,1,2,3,2,3,4,1,2,3;31,16,15,9,10,8,13,6,16,11,4,9,4,10,6,17,15];
A= [1,1,2,2,2;1,2,4,1,2;NaN,NaN,NaN,NaN,NaN;31,13,16,10,17];
donotMatch = zeros(size(A));


% find rows containing NaN
for col = 1:1:size(A,2)
    for row = 1:1:size(A,1)
        if (isequal(A(row,col), M(row,col)))
            continue
        elseif isequaln(A(row,col), NaN)
            donotMatch(row,col) = row;
        end
    end
end
disp(num2str(donotMatch))

% match only zero containing rows column by column
% which means ignore NaN rows
% How to know which rows to ignore while matching?
% That is why I have created donotMatch matrix
for col1 = 1:1:size(A,2)
   for col2 = 1:1:size(M,2)
       S = donotMatch(:,col1);
       if isequal(A(find(S==0),col1), M(find(S==0),col2))
           R(:,col1) = M(:,col2);
           break
       end       
   end
end
disp(num2str(R))

[Awais Saeed]

I figured out that first For Loop is not required. You can create donotMatch matrix as

donotMatch = isnan(A);

[Radu Mihail]

Yes Awais

Your method works. Great!

[Radu Mihail]

Thank you very much!

[Awais Saeed]

You are welcome.