Changing empty array for peak width to zero

So I have a matrix full of 25 iterations of a code (ends up as 75x25). For each iteration I wanted to calculate the width of the peak thats created, and I did this by using findpeaks. The problem is that my resulting width vector only includes 10 widths, and I'm assuming its because the other iterations do not have a peak, so the width is an empty array []. Is there a way I can insert zeros when they occur?

This is what I've tried

for i=1:25
[pks1(i,1), loc1(i,1), width1(i,1), prom1(i,1)]= findpeaks(XmRNA(:,i),t);
   for i=1:length(width1)
 if isempty(width1(i))
       width1(i)= 0
 end
   end
end

I tried using is empty but my resulting width vector doesnt include the empty arrays ([]) so it doesnt work. (To clarify my width1 vector is 10 numbers, so there are no empty arrays to change to zero).

Any help is welcome!

Answer 1

[Sindar]

% preallocate default values
pks1 = nan(25,1);
loc1 = nan(25,1);
width1 = zeros(25,1);
prom1 = nan(25,1);
for i=1:25
    [tmp_pk, tmp_loc, tmp_width, tmp_prom]= findpeaks(XmRNA(:,i),t);
    % if peak is found, update values with the *first* peak
    if ~isempty(pks)
        pks(i) = tmp_pk(1);
        loc1(i) = tmp_loc(1);
        width1(i) = tmp_width(1);
        prom1(i) = tmp_prom(1);
    end
end

you should think about whether you want the first peak or a different one (e.g., the most prominent peak)

Note: may be worth checking speed of parfor