how can i build a for loop to create histograms for a multiple cell array?

i have a multiple array cells (result):

there are 20 cells and each cell contains other 20 cells. I would like to create an histogram of each cell data, so in total 400 histograms. I built this code but if I have ii=1 i have L_nod starting from result{1,1}{1,2}. I want that it starts from result{1,1}{1,1} and increase each time.

for ii = 1:length(result)
    figure(ii)
    L_nod = result{1,ii}(:,ii+1)
    L_nodo1 = cell2mat(L_nod);
    [n,x] = hist(L_nodo1,50);
    n = n/length(L_nodo1)/diff(x(1:2));
    bar(x,n,'hist')
    pngFileName = sprintf('hist%d.png',ii);
    fullFileName = fullfile(folder, pngFileName);
    saveas(gcf,fullFileName);
end

maybe is not clear, if anybody has a new code to suggest to me i'm glad :).

Thanks!

[Image Analyst]

Can you attach the results in a .mat file?

And why is result a cell array? From what I see, it could be done much simpler with just a regular double array. I mean, none of the arrays have different sizes -- they're all 80 by 1 -- so why use cells instead of a 2-D double array?

[EM geo]

yes sure, result.mat is now attached! there are on or more array with size 79 by one, that's why i didn't create the matrix. the results are from this loop, if you can suggest me a way to modify it to have a matrix instead of a cella array i will be glad!!!!! :D

i attached also the input file.

clc; clear; close all;
load ('sclerometrica_equotip_v1')
misure = [ID_stop Nn Bn L];
% misure(any(isnan(misure), 2), :) = [];

ID_stop1 = misure(:,1);
L1 = misure(:,4);
Nn1 = misure(:,2);
Bn1 = misure(:,3);

k = unique(ID_stop1);
for i = 1:numel(k)
    index = (ID_stop1 == k(i));
    nn = Nn1(index);
    bn = Bn1(index);
    ll = L1(index);
    j =unique(nn);
    for ii= 1:numel(j)
    index1 = (nn == j(ii));
    Lnodo = ll(index1);
    result{i}{ii} = Lnodo;
    end
end

really thanks!!!

Answer 1

[Guillaume]

First, don't use 2D indexing for vectors. result{ii} will work whether result is a row or column vector. Your result{1, ii} will only work for a row vector and fail if that untested assumption is broken.

Note that the output of result{ii} (or result{1, ii}) is, as you've stated a 1x20 cell array. Again, it's a vector and you're using 2D indexing on that with index (:, ii+1). Since that cell array has only 1 row, : is equivalent to 1, so your Lnodo1 is:

Lnodo1 = cell2mat(result{1, ii}(1, ii+1));

written a lot more simply as

Lnodo1 = result{ii}{ii+1}; %the use of cell2mat is pointless, {} indexing instead of () indexing yields the same result

So basically, you Lnodo1 is in turn result{1}{2}, result{2}{3}, ..., result{20}{21}. That last one will cause an index exceeds matrix dimension error as well. Certainly it doesn't do what you want but I'm not sure why you didn't spot that.

Since you want to interate over the cells of the outer array and the cells of the inner array, you need a double for loop

for outer = 1:numel(result)  %prefer numel to length, it's safer
    for inner = 1:numeL(result{outer})
        L_nodo1 = result{outer}{inner};
        %... your histogram code. Perfer histogram to hist
        pngfilename = sprintf('hist%d_%d.png', outer, inner);
        %... rest of export code
    end
end

As Image Analyst suggest, you could avoid the cell arrays entirely, and store your data as an 80x400 matrix, which would make your life simpler, would use a lot less memory and be faster to process.

[EM geo]

your answer is really clear!!! thank you a lot!!!!!