finding out Mahalanobis distance between two matrix

I have two matrix

G1 =
   80.00
   79.43
   78.89
   78.38
   77.90
   77.45
   77.03
   76.63

G2 =     NaN       NaN       NaN
         NaN       NaN       NaN
         NaN       79.11     NaN
         NaN       78.64     NaN
         NaN       78.13     78.14
         77.46     NaN       77.66
         77.07     77.08     NaN
         76.74     NaN       NaN

I want to have result matrix which has a size of 8*1. Moreover, the G2 matrix elements should be compared with G1 and the nearest elements (Using only Mahalonobis distance) of G2 to G1 shoud be filled up in desired matrix. for example first one of G2 row has NaN, so it keep it NaN. 4th row of G2 has two values, this values should be compared with 4th element of G1 and nearest value shoud be filled up using Mahalanobis distance in the result matrix. When there is a single value than it shoud be remain like this only. When there are more than two values than the nearest value using mahalanobis distance shoud be filled up in the matrix.

Below is the desired output. How to do it more dynamically. i just did it by manually.

Result =   NaN
             NaN
             79.11
             78.64
             78.14
             77.46
             77.07
             76.744

i have already the solution for nearest neighbour way but i want it using Mahalanobis distance. When i tried using Mahal function than the error is that coloums should be same in both matrixs.

Answer 1

[VM Sreeram]

I am not sure why you got the error unless you share your implementation. You might have missed the transpose in line 9 of my implementaion.

G1 = [80.00; 79.43; 78.89; 78.38; 77.90; 77.45; 77.03; 76.63];
G2 = [NaN NaN NaN; NaN NaN NaN; NaN 79.11 NaN; NaN 78.64 NaN; NaN 78.13 78.14; 77.46 NaN 77.66; 77.07 77.08 NaN; 76.74 NaN NaN];

Result = nan(size(G1));
for i = 1:size(G2,1)
    row = G2(i,:);
    row = row(~isnan(row));
    if ~isempty(row)
        d = mahal(row', G1(i));    % row'
        [~, idx] = min(d);
        Result(i) = row(idx);
    end
end
Result

Result = 8×1

NaN NaN 79.1100 78.6400 78.1300 77.4600 77.0700 76.7400