When copying plot from one subplot position to another using copyobj, the second plot exhibits "hold on" behavior while the first doesn't.

Here's a bare bones version of the code that illustrates the problem. I want both plots to not hold on to previous iterations. While the plot on the top does not hold on to previous plots, the zoomed in plot does. I have even tried putting "hold off;" after every line following subplot(2,1,2) and it does not help.

clear; close all; clc;

counter = 0;
r = 0;
position = [0,0];
while counter < 100
    subplot(2,1,1)
    fimplicit(@(x,y) (x-position(1)).^2+(y-position(2)).^2-r.^2,'r')
    r = r + 0.1; position = position + [0.1,0];
    ax = gca; axis equal; axis([-10 10 -10 10]);

    zoomed = subplot(2,1,2);
    copyobj(get(ax,'Children'),zoomed)
    axis equal; axis([-1 1 -1 1]);
    pause(0.1)
    counter = counter + 1;
end

Answer 1

[Walter Roberson]

None of the low-level graphics operations pay attention to the hold state; they just add to (or modify) what is there already. The hold state is only paid attention to by the high-level plotting operations such as plot() and surf(), which contain code that specifically detects the hold state and clears the axes if the hold is not active.

When you copyobj() you are not invoking any high-level graphics operations, so the object is just added to whatever is there already.

You should cla() as needed.

You should also re-write the code so that every graphics operation is specific about the object it is going to act on.

    ax = subplot(2,1,1)
    fimplicit(ax, @(x,y) (x-position(1)).^2+(y-position(2)).^2-r.^2,'r')
    r = r + 0.1; position = position + [0.1,0];
    axis(ax, 'equal'); axis(ax, [-10 10 -10 10]);
    zoomed = subplot(2,1,2);
    cla(zoomed);
    copyobj(get(ax,'Children'),zoomed)
    axis(zoomed, 'equal'); axis(zoomed, [-1 1 -1 1]);
    pause(0.1)
    counter = counter + 1;

[Jacob Ward]

Thanks! Makes sense that all that was needed was to clear the zoomed axes of what was there before.

Answer 2

[Image Analyst]

You never called hold on for the top plot so why should it hold onto previously plotted stuff?

[Jacob Ward]

Exactly, I do not want either one to hold on. I never called hold on anywhere, but yet the second plot is holding on.