Matlab plot of categorical x-axis and cell array data

J Paul :

I have a cell array with nested cell arrays:

I want to plot each nested cell array by rows. But not all nested cell arrays are 8x1. I need to fill in the empy values as NaN or zeros, but still be able to plot the data continously.

example for columns 7-9:

Column7   Column8   Column9
1          1           1                       
2          2           2                        
3        NaN          NaN
4        NaN          NaN
5        NaN          NaN
6        NaN          NaN

I want to plot by rows, row 1 is (1,1,1), row 2 is (2,2,2), row 3 is (3, NaN, NaN), and so on; So category 7 will have values 1-6 vertically. Category 8 will have values 1 to NaN vertically but only plotting values 1 and 2.

I want row 1 values to connect via line:

Example:

figure
hold on
cellfun(@(C1) plot(cell2mat(C1,:), 'o-'), C);
% setup axes
xlim([0, 15]);
ax = gca;
ax.XTick = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15];
ax.XTickLabel = {'1','2','3','4','5','6','7','8','9','10','11','12','13','14','15'};

The plot should look like this:

IvanA :

The easiest way to do this, is to build a 2D matrix that contains your data (where present), and NaNs elsewhere. Most plotting commands in Matlab ignore NaNs.

To convert your cell array to a matrix, you can create a matrix containing only NaNs, and fill in the data present column-by-column. For example;

% Create dummy data
C = {[1;2;3;4], [7;7;8], [5;4]};
% Find the maximum number of rows possible across all cells
Nrows = max(cellfun(@length, C));
% Create matrix full of NaNs
M = nan(Nrows, length(C));
% Loop cells
for i = 1 : length(C)
    % Pull the contents of this cell
    Content = C{i};    
    % Fill this column with as many rows as we found
    M(1 : length(Content), i) = Content;
end

The resulting matrix M contains the values of each cell in C, one per column. You can then simply plot them with