how to combine 2 matrix at once with loop process..?

I have a hard case (in my opinion), and i need your favor for this thing..

I have a data like below :

data = [1   0
        2   0
        3   0
        4   0
        5   0
        6   0
        7   0
        8   0
        9   0
        10  0
        11  0
        12  0
        13  0
        14  0
        15  0
        16  0
        17  0
        18  0
        19  0
        20  0];

And 2 matrix like below :

A =

    11     3
    17     1
     5    19
     6    18
     6    16
    18     7
    11    15
    14    13
     4     2
    20     4

B =

    4.0078    1.9147
    4.9880    1.8967
    1.7225    1.1802
    1.1200    3.1427
    2.6425    4.9374
    2.4181    4.8850
    3.6921    2.9100
    3.0406    4.6255
    3.5859    4.3326
    2.4275    3.6506

I want to change the value in matrix 'data' row 2 with matrix B, the changes according to the matrix A. Matrix A use as a number that will synchronize with matrix 'data' row 1. For example, let see the first line of matrix A and matrix B. In matrix 'data' row 2 the value in line 11 and 3 (elements of matrix A) will change from 0 to 4.0078 and 1.9147 (elements of matrix B), so the matrix data will change like below :

data = [1   0
        2   0
        3   1.9147
        4   0
        5   0
        6   0
        7   0
        8   0
        9   0
        10  0
        11  4.0078
        12  0
        13  0
        14  0
        15  0
        16  0
        17  0
        18  0
        19  0
        20  0     ];

The changes will looping per line of matrix A and B, so the results will generates 10 different matrix data combination. I already tried to coding it, but there’s still wrong and I stuck with it. My coding shown below :

A = zeros(10,2);
B = zeros(10,2);

 for C = 1:10,

    d = 20;
    a = randperm(d);
    a = a(1:2);
    A(C,:) = a;

    b = (1+(5-1).*rand(1,2));
    B(C,:) = b;

 end

 for C = 1:10
    for i1 = 1:size(a),
        for i2 = 1:size(b),
            index = a(i1);
            if data(index, 2) == 0
               data(index, 2) = b(i2);
            end
        end
    end
 end

Did anyone had the solution..?

Thank you..

[Matt Kindig]

How do you handle the case where A has repeated indices, but B does not have repeated values? For example, in your shown data, the A(4,1) and A(5,1) elements both =4, but B(4,1) ~= B(5,1).

[Noru]

in my case above there's no problem with that, because line 4 and line 5 will applied in different matrix 'data'.

the problem is how to insert elements of matrix A and B per line to matrix 'data' so it will create 10 different matrix 'data' combination. one line for one matrix 'data'.

do you have any solutions..?

Answer 1

[Andrei Bobrov]

sa = size(A,1);
out = accumarray([A(:), repmat((1:sa)',2,1)],B(:),[size(data,1) sa]);

or

sa = size(A,1);
sd = size(data,1);
out = zeros(sd,sa);
out(sub2ind([sd,sa],A,repmat(1:sa,2,1)')) = B;

ADD

sd = size(data,1);
data = reshape(num2cell(permute(cat(3,repmat((1:sd)',1,sa),out),...
                                                             [1 3 2]),[1 2]),[],1);

[Noru]

thank you for your help, your solution real close to my expectation but i want to separate each combination of matrix 'data' with loop operation, so there are 10 matrix 'data' and not merged into one matrix 'data'..

how to do that..?

thank you..

[Andrei Bobrov]

see part ADD in my answer.

[Noru]

ok thank you sir...