Separating a vector based on its value

I currently have a vector of angle data (0-360). I am trying to create a loop that will create a new vector each time a value lower than 3 occurs as my data, as it has more than on 'cycles' in it. The code I use currently is very clunky, and I am sure there is a more simple loop that could do this:

x = find(diff(CA>352));
xx = x(2:2:end);
CA1 = CA(1:(xx(1)));
CA2 = CA((xx(1)+1):(xx(2)));
CA3 = CA((xx(2)+1):(xx(3)));
CA4 = CA((xx(3)+1):(xx(4)));
CA5 = CA((xx(4)+1):(xx(5)));

Any advise or solution would be great!

[Azzi Abdelmalek]

Why diff(CA>352) ?

[Jessica Smith]

I used this as the recorded angle didn't always hit '0' or '360', therefore, using this number seemed to identify a big enough change in angle to know a new cycle had begun. Again, admittedly very clunky

[Image Analyst]

What I did in my answer was to "create a new vector each time a value lower than 3 occurs" - did you see it. I also have no idea why 352 was used, so I used 3 like you said in the quote I just gave.

Answer 1

[Azzi Abdelmalek]

v=randi([0 360],1,1000)
threshold=3
out=accumarray((cumsum(v<=threshold)+1*~(v(1)<=threshold))',1:numel(v),[],@(x) {v(x)})

Answer 2

[Image Analyst]

If you have the Image Processing Toolbox, you can do it (find all contiguous runs of data less than value of 3) all in just 2 lines of code.

% Create random data
data = 10 * rand(1,100)

% Now, here is the main code.
% Find places where data is less than 3
lessThan3 = data < 3;
% Get coordinates of those into new vectors.
% Each "run" of data less than 3 is in a vector called PixelValues
% that is a field of a structure array called measurements
measurements = regionprops(lessThan3, data, 'PixelValues');
% NOW WE'RE ALL DONE!

% For fun, type out all those vectors
for k = 1 : length(measurements)
  fprintf('For run #%d, there are %d elements : ', k, length(measurements(k).PixelValues));
  fprintf('%.3f ', measurements(k).PixelValues);
  fprintf('\n');
end

Here's the data it found:

For run #1, there are 1 elements : 1.898 
For run #2, there are 4 elements : 2.414 1.204 2.450 2.037 
For run #3, there are 1 elements : 2.415 
For run #4, there are 1 elements : 1.535 
For run #5, there are 3 elements : 1.942 2.866 0.586 
For run #6, there are 1 elements : 2.383 
For run #7, there are 1 elements : 1.805 
For run #8, there are 2 elements : 1.465 1.349 
For run #9, there are 1 elements : 2.462 
For run #10, there are 1 elements : 1.067 
For run #11, there are 2 elements : 1.888 2.140 
For run #12, there are 1 elements : 0.324 
For run #13, there are 2 elements : 2.392 1.706 
For run #14, there are 2 elements : 0.594 1.408 
For run #15, there are 1 elements : 0.563 
For run #16, there are 1 elements : 2.259 
For run #17, there are 3 elements : 1.324 1.736 2.766 
For run #18, there are 1 elements : 0.917 
For run #19, there are 1 elements : 1.377 
For run #20, there are 1 elements : 2.709 
For run #21, there are 1 elements : 1.516

You can even do it all in one line of code if you want:

measurements = regionprops(data<3, data, 'PixelValues');