How do I extract data from a specific subplot in a MATLAB figure?

I have a matlab figure with 12 subplots (6X2 figures) and I intend to convert certain subplots into data (whose position I know). I have seen tutorials where MATLAB figures where converted to data but could not find one that handled extraction of data from subplots.

I am new to matlab and am having difficulties figuring out how to do this.

Any help is much appreciated!

Answer 1

[Luna]

Hi,

If you don't have the handles of the plot in the figure here is a demonstration for you.

Please read the comments carefully.

%% demo data preallocation
n = zeros(1,12);
x = cell(12,1);
y = cell(12,1);
hFig = figure;

% randomly creating x and y data and plotting
for i = 1:12
    n(i) = randi(20,1,1);
    x{i} = rand(n(i),1);
    y{i} = rand(n(i),1);   
    hAxes{i} = subplot(6,2,i); % hAxes axis object handles in a cell array (12 handles you have)
    hPlot{i} = plot(hAxes{i},x{i},y{i},'LineStyle','none','Marker','*'); % hPlot plot line object handles (12 handles you have)
end
% -------------------------------------------%

% Gathering data from plotted figure
%% FIRST OPTION IN CASE YOU HAVE HANDLES 
%% If you know the position use this: (That I choose the 6th position)
myPlotHandle =  hPlot{6};
xdata = myPlotHandle.XData;
ydata = myPlotHandle.YData;

% You can also use this from the position: (Axes' children is a plot line object. Our hPlot{6} is hAxes{6}.Children )
xdata = hAxes{6}.Children.XData;
ydata = hAxes{6}.Children.YData;

%%% OR %%%

%% SECOND OPTION IN CASE YOU DON'T HAVE THE OBJECT HANDLES 
% click to a point or a data line in the axis. You should click on a data
% plotted. 
% run the code below: (gco function gets you the current object(that means last focused or last clicked object), in this case that will be your plot handle like above)
myPlotHandle = gco;
xdata = obj.XData;
ydata = obj.YData;

% you can click for each 12 plot data and run the code so you will get them
% one by one. 

[Harishankar Anil Karingathuruthil]

Thank you, I tried with my file and it works.

Thank you very much!

Also FYI I used a different code to get the results from an alreaady existing MATLAB figure.

h=openfig('your_file_name.fig');
handles=findobj(h,'Type','line');
x=get(handles(1),'Xdata');
y=get(handles(1),'Ydata');

Here 1 is taken as the index of the subplot.

Cheers!

[Luna]

Your welcome :) Please accept answer if it works, thanks :)

Answer 2

[per isakson]

This illustrates the old way to do it. (There might be shortcut nowadays.)

%%  Example from the documentation of subplot
subplot(2,1,1);
x = linspace(0,10);
y1 = sin(x);
plot(x,y1)
%
subplot(2,1,2); 
y2 = sin(5*x);
plot(x,y2)
%%  Recover x, y1 and y2
ohf = findobj(gcf);      % gcf get current figure
%%
oha = findobj( ohf(1), 'Type','axes' ); % find the two axes
%%
oh1 = findobj( oha(1), 'Type','line' ); % find the line of the first axes
%%
all( oh1.XData == x )        % check if the data of the plot is equal to 
all( oh1.YData == y2 )       % the data plotted
%%
oh2 = findobj( oha(2), 'Type','line' ); 
%%
all( oh2.XData == x ) 
all( oh2.YData == y1 ) 

Which axes corresponds to which plot? The Position helps to find out

>> oha(1).Position
ans =
         0.13         0.11        0.775      0.34116
>> oha(2).Position
ans =
         0.13      0.58384        0.775      0.34116
>> 

oha(2) is the top one

ADDENDUM

Don't reinvent the wheel before checking with the File Exchange. (as I just did).

[per isakson]

ohf = findobj(gcf);      % gcf get current figure
%%
oha = findobj( ohf(1), 'Type','axes' );

may be replaced by a better way

oha = findobj( gcf, 'Type','axes' );

The figures can alternatively be found by

>> ohf = findobj( 0, 'Type', 'figure' );
>> ohf
ohf = 
  Figure (1) with properties:

      Number: 1
        Name: ''
       Color: [0.9400 0.9400 0.9400]
    Position: [680 678 560 420]
       Units: 'pixels'
       

With many figures one has to find a way to distinguish between them.

It's more robust not to use gcf, gca, gco in code.

[Harishankar Anil Karingathuruthil]

Thank you for the codes,

FYI I also found another method too, please check the comment in the above post!