CODEWITHSUNDEEP
matlabplotdata-visualizationline-plot
Kev1n91
Kev1n91

Reputation: 3693

How to ignore values that are negative when plotting a line plot with matlab?

I have a matrix where some values have -1 as an indication that there was an error. Normally I would just use ylim([0 100]) to not show these values in my graph, but when using line graph the connection will still drop down to the point. I want a chart that consists of lines, not a scatter plot. Is there a simple way of ignoring negative values when plotting a line and only connect positive values when using the plot function in MATLAB?

I have written a small example program which behaves similarly, but the way I use seems a little "too complicated" and I want to know if there is an easier way to achieve this. It works fine as I put the values to NaN, and now the x and y values are the same amount. However deleting or sorting out the values from the vector will lead to different amount of x and y values.

I was hoping for a modification or a flag or something.

x = 2*rand(10) - rand(10)
xx = 10:10:100;

figure;
for i=1:length(x)
    for j=1:length(x(i,:))
        if x(i,j) < 0
           x(i,j) = NaN; 
        end
    end
end

plot(xx,x)

Please note that this is only an example, the whole code would be too large to post here.

When having non-corresponding x-values (so that the plot function simply uses 1,2,3... and so on for the corresponding y values) this can be achieved by using 

plot(x(x>0))

In this case, the corresponding values are different, in the real code they are measured data, here I simply use a 10th step for simplification. 

x = 2*rand(10) - rand(10)
xx = 10:10:100;

plot(xx,x(x>0))

The above code will error with the message "Vectors must be the same length".

Upvotes: 1

Views: 4127

Answers (2)

Dev-iL
Dev-iL

Reputation: 24179

This is just like using NaN instead of negative values, only that the original vector is not modified at all. You might notice that this solution is vectorized.

y = 2*randn(10,1) - randn(10,1);
figure(); plot(1:numel(y), y./(y<=0) );

Upvotes: 4

Tony Tannous
Tony Tannous

Reputation: 14863

a = [50, -1, 10, 5, 8, 22, -1];
b = a > 0;
c = a(b);

Output:

c
[50, 10, 5, 8, 22]

Now you can plot c

[~, s] = size(c);
xx = 1:1:s;

You could also do it directly without saving the calculation and modifying it. just plot it. plot(x(x>0))

Upvotes: 1

Related Questions