placement of elements from a matrix into another matrix

Hi,

I have A matrix like below.

tot_row=5;
tot_col=10;
left_col_begin = tot_col/2;
right_col_begin = left_col_begin + 1;
new_col=3;
left_col_end =  (left_col_begin - new_col)+1;
right_col_end = (right_col_begin + new_col)-1;
new_row=3;
A=nan(tot_row,tot_col); 

And I have B matrix like below:

B=[10 5
6     4
3     2
8     0
1     3
7     5];

Now I want to get C matrix from A and B such that

B(i,1) for all i= 1:size(B,1)

will be placed on A matrix based on their frequency given in B(i,2)

But for each row rr in A matrix where rr in 1:new_row --

That means for above A matrix, the placement of B(i,1) will be B(i,2) number of times in the following place sequence

A(1,6) -> A(1,5) -> A(1,7) -> A(1,4) -> A(1,8) -> A(1,3)
A(2,6) -> A(2,5) -> A(2,7) -> A(2,4) -> A(2,8) -> A(2,3)
A(3,6) -> A(3,5) -> A(3,7) -> A(3,4) -> A(3,8) -> A(3,3)

My output matrix C is given below. I want this C matrix from A and B based on above placement criteria.

C=
[NaN      NaN       6         10        10        10        10        10        NaN       NaN
NaN       NaN       1          3         6         6        6         3         NaN       NaN
NaN       NaN       7          7         1         1        7         7         NaN       NaN
NaN       NaN       NaN       NaN       NaN       NaN       NaN       NaN       NaN       NaN
NaN       NaN       NaN       NaN       NaN       NaN       NaN       NaN       NaN       NaN];

7 should have 5 times in C matrix as it's frequency is 5 in B matrix, but it is here available for 4 times. Because I don’t have any other spaces (in that 3*6 block) to put 7. So this is fine. My goal is to place as much as I can. If there are any extra space available in that block after all the placements from B matrix, then it will be keep as NaN as like rest of the elements.

Can someone please help me on this? I hope I made my question clear. Please let me know if you have any questions.

Thanks

Answer 1

[Andrei Bobrov]

%{ 
block of code by Rupsana - initial and final submatrix values and initial
matrix A:
%} 
tot_row=5;
tot_col=10;
left_col_begin = tot_col/2;
right_col_begin = left_col_begin + 1;
new_col=3;
left_col_end =  (left_col_begin - new_col) + 1;
right_col_end = (right_col_begin + new_col) - 1;
new_row=3;
A=nan(tot_row,tot_col); 

% ii - sequence of columns in submatrix
ii = reshape(...
   [right_col_begin : sign(right_col_end - right_col_begin) : right_col_end
   left_col_begin : sign(left_col_end - left_col_begin) : left_col_end]   ,[],1);
[y,x] = ndgrid(1:new_row,ii);

% adjustment of the number of elements for the submatrix
k = cumsum(B(:,2),'reverse');
d = mod(k(1),numel(ii)*new_row);
jj = find(k-d > 0,1,'last');
B(jj,2) = B(jj,2) - d;
B(jj+1:end,2) = 0;

% submatrix insertion into matrix "A"
A(reshape(sub2ind([tot_row,tot_col],y,x)',[],1)) = repelem(B(:,1),B(:,2));