How can i sort data using logical indexing?

problem: i have two arrays: the first one is of unknown lenght and has two columns, one containing the data that i want to add to my second array and one containing a natural number (ranging from 1 to unknown) looking like this:

[                   0]    [3]

[                   0]    [5]

[                3.36]    [ 2]

[ 0.00100000000000122]    [ 6]

[0.000999999999999446]    [ 4]

[0.000999999999999446]    [ 2]

my second array is empty. now i want to add the values (first columns) to the associated row ( second column) in my second array. There are multiple values for each row. the output of the second array should look like this:

[] []

[3.36] [0.00099999999446]

[0] []

[0.0009999999999446] []

[0] []

[0.00100000000000122] []

Is there a way to do this using logical indexing? It would be nice if i could avoid using for loops with if statements

Thanks.

[the cyclist]

Can you write explicitly what you want the output to be, for this example?

[Timo Merker]

sorry i hit the enter button before i was done. added the output!

Thanks

Answer 1

[Guillaume]

Redistributing your numbers in the row indicated by the second column is very easy with accumarray:

a = [0                        3
     0                        5
     3.36000000000000         2
     0.00100000000000122      6
     0.000999999999999446     4
     0.000999999999999446     2];
b = accumarray(a(:, 2), a(:, 1), [], @(v) {v.'})

This creates a 1d (column) cell array where each row is all the numbers corresponding to that particular row index. Each row is just long enough to store the required numbers

The problem comes from you wanting a 2d cell array instead. To be honest, I'm not sure there's any advantage in it and it requires more work. One way of generating it:

%continuing from above
maxcols = max(cellfun(@numel, b)); %or you could get that from the histogram of a(:, 2)
c = cellfun(@(v) [num2cell(v), cell(1, maxcols - numel(v))], b, 'UniformOutput', false);
c = vertcat(c{:})

[Timo Merker]

I tried using the first option but i get this error: Error using accumarray Second input VAL must be a vector with the same number of elements as each cell in SUBS, or a scalar.

can this be caused by empty cells in column 1?

[Guillaume]

It's important to use valid syntax in your question to avoid confusion. From your description, I assumed that the input array is a standard array (i.e. matrix), not a cell array.

Now, why is your input a cell array? and more importantly, why does it have empty cells in the first column?

If the cell is empty, it's not going to contribute anything to the output, so you might as well remove that row. Once the input cell array has no empty rows, you can just convert it to a matrix and my answer will work:

a = {0                        3
     0                        5
     []                       3
     3.36000000000000         2
     0.00100000000000122      6
     0.000999999999999446     4
     0.000999999999999446     2}; %a is now a cell array not a matrix
 a(any(cellfun(@isempty, a), 2), :) = []; %remove rows with empty values
 a = cell2mat(a)
 %now you can use accumarray

[Timo Merker]

you are right, i didnt use the correct syntax. I'm new to matlab so im still struggeling with this. Im dealing with a huge dataset and though it should not contain empty cells i wasnt shure. checked it, fixed it, converted to matrix like you said and it works perfectly! So many thanks for your time and patience, i appreciate it! :-)