error subscript indices integers or logicals

for this code i get the error shown above, why? in particular it is for l = numel(((wechsel_X_idx(v):wechsel_X_idx(v+1))/2)-1); Y(l) = meaned_Y; i dont get it because when i try it with numbers without a for loop it works.. thanks anybody who can help!

if true
wechsel_X = diff(wertepaare(:,2));
wechsel_X_idx = find(wechsel_X);
  for v = 1:(numel(wechsel_X_idx)-1)
      meaned_Y = mean(Y(wechsel_X_idx(v)+1:wechsel_X_idx(v+1)));
      Y(wechsel_X_idx(v):wechsel_X_idx(v+1)) = 0;
      l = numel(((wechsel_X_idx(v):wechsel_X_idx(v+1))/2)-1);
      Y(l) = meaned_Y;
  end  end

[Alexandra Harkai]

Which v value are you getting the error for? (Seems unlikely that line would cause the problem since the same indexing works in the line before.)

Answer 1

[Guillaume]

Without knowing the content of the variables, it's difficult for us to diagnose the problem. However, the line:

l = numel(((wechsel_X_idx(v):wechsel_X_idx(v+1))/2)-1);

is very suspicious. You're asking with numel how many elements there are in a vector. That vector is made of the values wechsel_X_idx(v):wechsel_X_idx(v+1). That you're dividing the content of the vector by 2 and then subtracting 1 to the elements is not going to change the number of elements, so your line is exactly the same as:

l = numel(wechsel_X_idx(v):wechsel_X_idx(v+1));

which would be obtained more simply and explicitly as:

l = wechsel_X_idx(v+1) -  wechsel_X_idx(v) + 1;  %note that wechsel_X_idx(v+1) is guaranteed to be greater than wechsel_X_idx(v)

So is the line doing what you want?

Note that the best way for you to find where the problem lies in your code is to debug it, executing each line step by step and inspecting the contents of the variable. It should quickly become obvious when the content deviates from your expectations.

[Payjay]

you are of course right with the "numel-thing", thank you! and no it didnt work as i wanted! now it works, think i mixed up some loops in an inappropriate way. and ur hint with the numel thing was also good, anyway thanks! :)