find possible combinations of each row of matrix and allocate their respective values into 2 different matrices

Hi everyone, I have 2 matrices, A and B. I would like to produce C and D by 1) listing down possible combinations within each row. If there is zero, just ignore it. For example, in row 1 of B, it will be just 2 3 while the row 2 will have combinations of 2 3, 2 4 and 3 4. then 2) I would like to put the values of A respectively to B's combinations

A = [0.3939
     0.3116]
 
B = [2 3 0 
     2 3 4]

expected results

 C = [0.3939 % from row 1 of A
      0.3116 % from row 2 of A
      0.3116 % from row 2 of A
      0.3116 % from row 2 of A
      ];
  
  
D = [2 3 % from row 1 of B
     2 3 % from row 2 of B
     2 4 % from row 2 of B
     3 5 % from row 2 of B
     ];

[the cyclist]

Should the last row of D actually be [3 4]?

How do you want to handle repeated elements in a row of B? For example, what if

B = [2 3 0 
     2 3 3]

[JL]

there's no repeat of the same number

Answer 1

[the cyclist]

% The input data
A = [0.3939
     0.3116];

 B = [1	16	17	6	9	0	0	0	0	0	0	0	0	0
      1	16	17	6	7	8	10	0	0	0	0	0	0	0
];

% Find the size of B, which is used a lot in the algorithm
[mb,nb] = size(B);
% Preallocate the output arrays, allowing for the most possible pairs
% within a row
maxPossiblePairs = nchoosek(nb,2);
C = zeros(mb*maxPossiblePairs,1);
D = zeros(mb*maxPossiblePairs,2);
% For each row of B, find the unique pairs of values. (Worry about zeros later.)
% Place those values into the correct segment of D.
% Put that row's value of A into the correct segment of C.
for ib = 1:mb
    rowIndex = (ib-1)*maxPossiblePairs+1:ib*maxPossiblePairs;
    D(rowIndex,:) = nchoosek(B(ib,:),2);
    C(rowIndex,:) = A(ib);
end
% Find and remove any rows with zeros in D
idxToRemove = any(D==0,2);
C(idxToRemove,:) = [];
D(idxToRemove,:) = [];

[JL]

Hi Cyclist, thanks for the code. How do I make it work if I have more columns in B matrix?

B = [1	16	17	6	9	0	0	0	0	0	0	0	0	0
1	16	17	6	7	8	10	0	0	0	0	0	0	0
];

[the cyclist]

I messed up the definition of the row indexing. I corrected the code above.

Also, the code will use memory more efficiently if you remove any rows of B in which the whole column is zero.

[JL]

Thank you very much! :)