MATLAB: Use cells rather than matrices to sort a matrix into 255 separate lists

cell arraysmatrix manipulationsort

I am trying to write a program that will sort an nx5 matrix into 255 separate lists (and it was suggested that I use cells to do this). n is usually upwards of 600,000. The numbers in the 5th column range from 1 to 255 (all are repeated), and they identify which basket the row belongs in. I am unfamiliar with how cells work, but since their size does not need to be pre-allocated, someone said this would be a good tool to use. I would like to label the baskets from 1 to 255, basket(i), or something to that effect, but I don't know the correct syntax to do that. "flag" and "switch" were also suggested. Would it be best to loop through the entire nx5 matrix and sort the rows into baskets that way? Any help with this would be appreciated. Thank you!

Best Answer

If I understand your question, the following may help

% create array like yours... n-by-5, last column has index of "bin" which
% will be in the set {1,2,...,maxIndex}
n = 10;
maxIndex = 5;
C = [rand(n,4),randi(maxIndex,n,1)];
% idx is a n-by-1 cell array where idx{j} contains the indices of rows
% belonging to bin j
idx = accumarray(C(:,5), 1:size(C,1),[],@(C){C});
% collect those rows from the matrix belonging to bin j into the jth cell
% of bins
bins = cellfun(@(x)C(x,:),idx,'uniformoutput',false);

Related Solutions

MATLAB: Manipulate data in cells

Hi Rebecca,

I think it would be easiest just to loop through all of the elements of the matrix. To access a given element in the cell array 'bins', use the curly braces {}, in a way that you similarly would use parentheses () for matrices.

For the first part:

 output = cell(size(bins));
 for k=1:length(bins), 
     mat = bins{k};   %get element in cell

     mat(:,5)=[];     %remove 5th column

     output{k} = mat;  %assign to output cell array 
end

The second part of your request is unclear. Once you break each row into four separate groups of three columns, what do you want to do with the data? Are you assigning each group as an element of the cell array (such that the final cell array is 255x4)? If so, this revised code will work:

 output = cell(length(bins), 4);
 for k=1:length(bins), 
     mat = bins{k};   %get element in cell
     mat(:,5)=[];     %remove 5th column
     output{k,1} = mat(:,[1 2 3]);  %break columns into groups and 
     output{k,2} = mat(:,[1 2 4]);  % assign to cell array
     output{k,3) = mat(:,[1 3 4]);
     output{k,4} = mat(:,[2 3 4]);
end

MATLAB: How to use histcounts with overlapping bins

Discretize three times per dimension, once with the bins exactly where you want them, once with the bins [overlap] earlier, once with the bins [overlap] later. Do the 27 different 3D binnings (each possible combination of early, middle, late), taking lists of indices. Then take the union of all of the indices in corresponding bins.

Best Answer

Related Solutions

MATLAB: Manipulate data in cells

MATLAB: How to use histcounts with overlapping bins

Related Question