Eliminate the values in the rows corresponding to a column which are zero in MATLAB

Question

I have a matrix of data which I am trying to analyse. I have a data and I applied some processing part and I managed to get some information below a certain level as in trying to apply threshold to it. So after I applied threshold the data goes to 0 point. So I was wondering if there was a way to just eliminate the points without it leaving 0 in between. This is what the figure looks like with the zeros in that I was trying to plot it without the gap the X Axis is time, y axis is amplitude. So will that be possible to just plot the events which are in blue and the time together?

%Find time
N = size(prcdata(:,1),1); 
t=T*(0:N-1)';
figure;
plot(t,U);
t1=t(1:length(t)/5);
X=(length(prcdata(:,4))/5);
a = U(1 : X);
threshold=3.063;
A=a>threshold;
plot_vals=a.*A;
figure; 
plot(t2,plot_vals1); %gives the plot which i added with this 

I also tried this code to club the events without the zeros but all it gives me is a straight line plot at 0.

%% Eliminate the rows and colomns which are zero
    B1=plot_vals1(plot_vals1 <= 0, :); 
    figure;
    plot(B1);

Also is there any way to take the scatter of the figure above? Will using scatter(t2,plot_vals1); work?

Answer 1

If you want to only display those points that are above your threshold, you can use a logical index and set the value of the unwanted points to NaN:

threshold = 3.063;
index = (a <= threshold);
a(index) = NaN;
figure;
plot(t1, a);

Data points that are NaN simply won't be displayed, leaving a break in your plot. Here's a simple example that plots the positive points of a sine wave in red:

t = linspace(0, 4*pi, 100);
y = sin(t);
plot(t, y)
hold on;
index = (y < 0);
y(index) = nan;
plot(t, y, 'r', 'LineWidth', 2);

Answer 2

So will that be possible to just plot the events which are in blue and the time together?

If the time of occurrence is not important to you, then the following will work:

After A=a>threshold;, change your code to

plot_vals=a(A);
figure; 
plot(plot_vals);

If the time of occurrence is important, then you can try setting the x-ticks and labels programmatically using the 'XTick' and 'XTickLabel' properties of your plot.

Get the corresponding time of interest like so:

t2=t1(A);

This should give you an idea of how to go about it using 5 equally spaced ticks:

xTickLabels = t2(floor(linspace(1,t2(end),5)));
xTicks = floor(linspace(1,numel(plot_vals),5));
plot(plot_vals);
set(gca,'XTick',xTicks,'XTickLabel',xTickLabels); % gca gets current axis handle

It is a bit of an art to determine which time points of interest you want, since your blue segments do not occur in equal sized clumps.