Plot segments of an array with NaN using a for loop

Hi,

I got an array with NaN. I want to plot, step-by-step, those segments of integers, that exceed a threshold and flanked by NaN, using a for loop.

Example:

Threshold=5
A=[NaN 1 2 3 4 4 3 1 NaN 2 4 5 10 11 12 14 13 11 7 5 3 NaN 2 3 7 15 20 25 17 13 9 5 1 NaN]

So, I want the loop to take the second segment of the integers flanked by NaN (2 - 3) and plot them. Then, I will do some operations with the plot, and I will continue with the loop using a wait(X) for the third segmentt (2 - 1).

The array always starts with a NaN, and finishes with an NaN.

for i = 1:length(A)
%     code for getting the plot and make the operation
wait(next_plot)
end

At least, if I can plot the second and the third segment together skipping the first one, I would save my day.

Best,

Jose

Answer 1

[dpb]

Rough outline to accomplish --

ix=find(isnan(A)); ix=ix(2:end);        % find the NaN sections, ignore first
for i=1:numel(ix)-1                     % over the groups
  subplot(2,1,i);                       % just to show the separate pieces;
  x=ix(i):ix(i+1);                      % set the x values to match
  B=A(x);                               % select the portion wanted from A this group
  B(B<=Threshold)=nan;                  % don't plot those <= Threshold
  plot(x,B,'*-')                        % plot verus original ordinal position
end

If only want the values above the threshold, then delete those below (or select only those above) instead and the corresponding x values to match.

[Jose Rego Terol]

Many thanks for your comment.

I am using the first lines of the code

ix=find(isnan(A)); ix=ix(2:end);        % find the NaN sections, ignore first
for i=1:numel(ix)-1

However, I am changing the rest because the rest of the code did not work for my code. I am usign ginput for each plot, therefore I cannot use subplots.

So, now I am doing like this but I got a problem to skip those segments that have lower values than the threshold (15)

ix=find(isnan(Trial(:,2))); ix=ix(1:end);        % find the NaN sections, ignore first
 for i=1:numel(ix)-1                    % numel(ix)-1 over the groups
  
  f=find(Trial((ix(1):ix(1+1)),2)>15); %Find all the values that are higher than 15 within this segment
  %%% Here I want to set the criteria: If f is empty, disp no spike found, and if f
  %%% has values, plot the entire segment (Trial)
  f=logical(f);
  if f == 0
      disp('no spike')
  else
      
      plot(Trial((ix(3):ix(3+1)),1),Trial((ix(3):ix(3+1)),2))
  end
pause()
%   plot(i);                       % just to show the separate pieces;
  
%   pause()
%   x=ix(i):ix(i+1);                      % set the x values to match
%   B=p(x);                               % select the portion wanted from A this group
%   B(B<=Threshold)=nan;                  % don't plot those <= Threshold
%   plot(x,B,'*-')                        % plot verus original ordinal position
end

The problem lies on the variable f. When the values are lower that 15, f is empty. Actually, I am looking for a code to say: in this segment or itineration, matlabdid not found values higher than 15. Then disp 'no spikes# and contiunue with the next segment.

[Jose Rego Terol]

Well, I got a workaround with numel()

for i=1:numel(ix)-1                    % numel(ix)-1 over the groups
  
   f=numel(find(Trial((ix(i):ix(i+1)),2)>15));
    
   if f == 0
       disp('no spike')
   else
       
       plot(Trial((ix(i):ix(i+1)),1),Trial((ix(i):ix(i+1)),2))
   end
 pause()
%    plot(i);                       % just to show the separate pieces;
  
%    pause()
%   x=ix(i):ix(i+1);                      % set the x values to match
%   B=Trial(x,2);                               % select the portion wanted from A this group
%   B(B<=Threshold)=nan;                  % don't plot those <= Threshold
%   plot(x,B)                        % plot verus original ordinal position
end

Now, if the segment does not have any value higher than 15, matlab disp 'no spike'.

Thanks for the first line. It push the code to do what I want to do.