i need to degenerate a matrix. I have 1 column with 11,2,2,2,3​,3..748,74​9,749 ( max 5 repetitions for each integer, but random, 2340 in total), I need to select the data in the rows with the same integer value and pick one at random to give 749 rows

i need to degenerate a matrix. I have 1 column with 1,1,2,2,2,3,3..748,749,749 ( max 5 repetitions for each integer, but random, 2340 in total), I need to select the data in the rows with the same integer value and pick one at random to give 749 rows

Answer 1

[Azzi Abdelmalek]

out=cell2mat(accumarray(A(:,1),(1:numel(A(:,1)))',[],@(x) {A(x(randi(numel(x))),:)}))

[ian benford]

Very efficient thankyou

Answer 2

[Stephen23]

Try this:

% Fake data:
A(:,2) = 0:9; % fake data from 0 to 9
A(:,1) = [1,1,1,2,2,2,2,3,3,3]; % group indices
% 
N = max(A(:,1));
for k = N:-1:1 % backwards is intentional!
    idx = find(A(:,1)==k);
    B(k,:) = A(idx(randi(numel(idx),1)),:);
end

And we can check the output in the command window:

>> A
A =
     1     0
     1     1
     1     2
     2     3
     2     4
     2     5
     2     6
     3     7
     3     8
     3     9
>> B
B =
     1     2
     2     3
     3     9

The rows in B are randomly selected from A, one for each group identified by the first column. Also note looping backwards is intentional, to preallocate the array B, and that using a loop is faster than using arrayfun or cellfun.

When I run the script a few times I get these outputs:

B =
     1     2
     2     4
     3     9
B =
     1     1
     2     4
     3     9
B =
     1     1
     2     3
     3     7
B =
     1     0
     2     6
     3     9