sorting 3d matrix by value of a cell

hei,

i have a 3d matrix, where each layer into the 3rd dimension is a new measurement. the value in one cell however gives me the distance. now i would like to sort these layer by this cell (distance).

this is possible within a 2d matrix with the function 'sortrows'.

thanks for help ;-) markus

Answer 1

[Laura Proctor]

From what I understand, you just want to sort by an element on each page in the 3D array - this is how you can do it:

A = randi(10,[4,4,4])
[~,idx] = sort(A,3);       % sort by the page
iv = zeros(1,size(A,3));   % set up the indexing vector
iv(:) = idx(1,1,:);        % find the ordering of the (1,1) 
                           % element by which to sort A
B = A(:,:,iv)              % sort A according to the (1,1)
                           % element of each page

[Markus]

hei laura,

thanks for your answer. if i read your code right, you sorted the pages by cell (1,1,:) which is the same position in A and idx.

;-) markus

[Laura Proctor]

You got it!

Answer 2

[Sven]

Hi Markus,

If you have a 3D MATRIX of numbers, and the distance you want to sort by is stored, say, in the (1,1) location of each slice in the third dimension, try:

A = rand(5,5,10);
[~,sortOrder] = sort(A(1,1,:),3);
A_sorted = A(:,:,sortOrder);

If you have a 3D CELL of numbers, read below:

Unfortunately when passing a cell to sort(), it expects that cell to be a cell of strings. The only way to sort numeric data in a cell is to convert it to a matrix, as below:

Let's first make your data into a cell (I'm only including "distance"... I presume since your data is in a cell to start with then there's a good reason you didn't store it as a vector of distances):

distanceCell = num2cell(rand(1,1,20))

We then just make a vector from only the numeric values in the cell

distancesVector = cat(1,distanceCell{1,1,:});

Then we sort that vector and reorder the cell accordingly.

[~, sortOrder] = sort(distancesVector);
sortedCell = distanceCell(:,:,sortOrder);

[Markus]

it is the first case. thats what i needed!! thanks..

[Markus]

aehm, i also accept your answer .. :) but i read laura first.

[Sven]

Aha, no problems Markus