How can I use sliders to do interactive plots from a structure array?

I have a structure array containing a B number of matrixes of [MxN] data, where M is different for every set and N is the same for every set of data.

I would like to have only one plot, where using two sliders I can choose between the B sets of data and N columns. I have only found examples of how to use a slider to update a plot changing a parameter (like the damping factor of a sinusoid), but nothing useful for my case.

[Stijn Haenen]

Something like this:

close all
clear
b=struct;
for i=1:10
b(i).data=ones(10)*i.*(i:i:i*10);       %creating struc with data
end

fig = figure;
s1 = uicontrol('Parent',fig,'Style','slider','Position',[81,1,419,23],...
              'value',0.1, 'min',0.1, 'max',10);
s2 = uicontrol('Parent',fig,'Style','slider','Position',[81,24,419,23],...
              'value',0.1, 'min',0.1, 'max',10);  %creating sliders
hold on
while 1>0
plot(b(ceil(s1.Value)).data(:,ceil(s2.Value)),'b')  %plot data
pause(0.1)
delete(fig.CurrentAxes.Children)
end

[Seth Meiselman]

Could you give an example of this structure array- how you've composed it in code? I think the solution above seems reasonable unless there is something specific about how you've constructed this data set. Are the matrices enumerated or labeled in the structure?

[Alessandro Giacetti]

A =

1×95 struct array with fields:

Batch

You can access the matrixes typing A(i).Batch with i = 1:95. The solution above works but it is a bit rough because it creates that infinite loop. I would like to use the callback instead.

[Rik]

If you write code using that callback (and I encourage you to try something yourself), make sure to round the value, as you're not guaranteed an integer.

Answer 1

[Seth Meiselman]

There's probably a better, cleaner way, but it should work.

clear all;
close all;

% Some junk data sets in a structure defined as A(b).data(m,n)
imax = 10;
jmax = 10;
datax = 0:pi/12:(23*pi);
for i=1:imax
    for j=1:jmax+i
        % Creating data struc A(b).data(m,n)
        % y = y0 + sin(w*x)
        blx(i).data(j,:)= i+sin((j/5)*datax); 
    end
    sizej(i) = size(blx(i).data(:,:),1);
end


% Plot different data sets according to slider location.
slider_plot(blx, datax, imax, jmax, sizej);


function [] = slider_plot(blx, datax, imax, jmax, sizej)
% Define figure window, normalized to fit arb. screen
S.fh = figure('units','normalized','name','slider_plot',...%%%%
              'Position', [0.515 0.025 0.415 0.87] );
% Define axes so that room is available in figure window for sliders
S.ax = axes('unit','normalized','position',[0.1 0.5 0.8 0.45]);
% There might be a better way to substitute data sets, but this works.
S.y = @(par) blx(par(1)).data(par(2),:);
% Define inital parameter values for input
S.a = 1;
S.b = 1;
% Plot 'junk data' as reference
S.p1 = scatter(datax,S.y([S.a,S.b]),'k');
hold on;
% Plot new data set on top of 'junk data' for visual comparison
S.p2 = plot(datax,S.y([S.a,S.b]));
hold off;
update(S);
% Slider for slope parameter:
S.aSlider = uicontrol('style','slider','unit','normalized',...
                      'position',[0.2 0.1 0.7 0.01],...
                      'Min',1,'Max',imax,'Value', 1,...
                      'sliderstep',[0.1 0.1],'callback', {@SliderCB, 'a',sizej}); 
% Add a text uicontrol to label the slider.
txta = uicontrol('Style','text','unit','normalized',...
                 'position',[0.2 0.11 0.7 0.02],...
                 'String','Data set "blx(i)"');
% 2nd Slider:
S.bSlider = uicontrol('style','slide','unit','normalized',...
                      'position',[0.2 0.15 0.7 0.01],...
                      'Min',1,'Max',sizej(S.a),'Value', 1,...
                      'sliderstep',[1/sizej(S.a) 1/sizej(S.a)],'callback', {@SliderCB, 'b', sizej});
% Add a text uicontrol to label the slider.
txtb = uicontrol('Style','text','unit','normalized',...
                 'position',[0.2 0.16 0.7 0.02],...
                 'String','Data line ".data(j,:)"');       
guidata(S.fh, S);  % Store S structure in the figure
end

% Callback for all sliders defined above
function SliderCB(Slider, EventData, Param, sizej)
%     S = guidata(Slider);  % Get S structure from the figure
%     val = round(get(Slider, 'value')); % round slider pos if manually adjusted
%     S.(Param) = val;  % Any of the 'a', 'b', etc. defined
%     update(S);  % Update the plot values
%     guidata(Slider, S);  % Store modified S in figure
%     
    if Param == 'a'
        S = guidata(Slider);  % Get S structure from the figure
        val = round(get(Slider, 'value')); % round slider pos if manually adjusted
        S.(Param) = val;  % Any of the 'a', 'b', etc. defined
        update(S);  % Update the plot values
        guidata(Slider, S);  % Store modified S in figure
        S.bSlider = uicontrol('style','slide','unit','normalized',...
                      'position',[0.2 0.15 0.7 0.01],...
                      'Min',1,'Max',sizej(val),'Value', 1,...
                      'sliderstep',[1/sizej(val) 1/sizej(val)],'callback', {@SliderCB, 'b', sizej});
    else
        S = guidata(Slider);  % Get S structure from the figure
        val = round(get(Slider, 'value')); % round slider pos if manually adjusted
        S.(Param) = val;  % Any of the 'a', 'b', etc. defined
        update(S);  % Update the plot values
        guidata(Slider, S);  % Store modified S in figure
    end
end

% Plot update, creates new y-vector for plot and replaces the plot
% S.p2 with new y-vector
function update(S)
yset = S.y([S.a,S.b]);   % General data set
set(S.p2, 'YData', yset);  % Replace old plot with new plotting values
end

[Alessandro Giacetti]

Ok, Thank you very much. Since both the number of the matrix and the number of column are integer, should I change the 'sliderstep' option in the uicontrol command?

[Seth Meiselman]

Try it first. Notice that the fake data I created at the top also has integer B and integer M.... all indices are integers in this script.

[Rik]

Why are you ignoring the mlint warning about clear all? You're overwriting all variables that might still exist from a previous run, so even clear isn't strictly necessary.

[Seth Meiselman]

Rik- as a stand alone script, I don't want it to interact with any leftovers in the workspace. You can remove those lines if you need to.

[Alessandro Giacetti]

Rik - Yes of course I removed those lines. I need to use this function inside another one.

Seth - The function works well, now I have a solid base where I might add some useful stuff. I would like to see something that tells me which matrix and which variable the plot show. I am going to work on that. Thank you very much

[Rik]

@Seth, you shouldn't be using 'clear all' in the first place. Use 'clear' or 'clearvars' instead.