Convert 1D array to 2D array with x and y coordinates supplied as individual vectors

Hi, I have 3 vectors: row=[1,3,2,2]; col=[2,1,2,1]; data=[1.8, 2.5, 3.9, 4.2]; I need to index the data into a 2D array using the row and column vectors stating the coordinates (with NaNs where I dont have any values) such that I get as my final answer as :

data2D =

    [  NaN    1.8000
    4.2000    3.9000
    2.5000       NaN]

The code that I am using now is as follows:

    Data2D = nan(max(row),max(col));
    for i =1:length (row);
        Data2D (row(i),col(i))= data(i); % Genarating Data Matrix %
    end

However, I am dealing with very large data sets and the for loop is proving to be quite slow. Is there anyway I can directly index the data with the row and col vectors without a for loop?

Answer 1

[Andrei Bobrov]

Data2D = accumarray([row(:),col(:)],data(:),[],[],nan);

[Himanish Basu]

Works Great! Thanks

[Guillaume]

Note that if you happen to have duplicate row column combination, this will sum the data for that point.

[Himanish Basu]

Can this be done if the data array is a cell array of strings, with everything else remaining the same? eg data={'A01','B03','D01','F12'}

[Guillaume]

It can't really be done with accumaray, but it can be done with sub2ind:

c = cell(max(row), max(col));
c(sub2ind([max(row) max(col)], row, col) = data;

Note that sub2ind will also work with numerical matrices.

[Jim Parsons]

@Guillaume - instead of having it take the sum, is there something out there that'll get it to take the average ?