Error in loop: Attempted to access X(5); index out of bounds because numel(X)=4

I tried searching, but don't seem to see any solutions for my problem. I have the entire code attached in a file, along with a valid data example.

In the following

for l = 1 : length(X)                        % uses the letter L
    if X(l) < 20 || X(l) > 200 || Y(l) < 50 || Y(l) > 500 ...
            || any(a{l} < 0) || any(a{l} > 360) ...
            || any(N{l} < 0) || any(N{l} > 50)
        X(l) = [];
        Y(l) = [];
        a{l} = [];
        N{l} = [];
        fprintf('Error in test subject %d, discarding dataline.\n',l)
    end
end

I get an error

Error in test subject 3, discarding dataline.

Attempted to access X(5); index out of bounds because numel(X)=4.

In this example there are datalines to go through. It 'deletes' each not-valid dataline, but unless it is the last dataline only that is discarded, this error will appear. Any way to solve this? The attached .m shows how i made the function (do tell if you want me to just c/p it here instead!).

Answer 1

[dpb]

Yes, a very common error in coding logic everybody makes when starting out...when you process from the beginning and remove a row, then you've just shortened the size of the array. Couple of ways to go at it; the simplest is simply to begin at the end and work forward...

for i=length(x):-1:1
  if condition, x(i)=[]; end
end

Another is to keep a list of the bad locations and wait until the end and do them all at once't...

ix=[];
for i=1:length(x)
  if condition, ix=[ix;i]; end
end
x(ix)=[];

Or, same idea, vectorize the test and eliminate the loop...

ix=logicaltest(x);  % return logical vector of lines to kill
x(ix)=[];

Or, of course, you can reverse the test logic and find those to keep instead.

[Kenan Hoyt]

dpb-superman, thanks! Noted all that down. Is there a list of "common errors" that appear frequently, that i can take a look at before asking the (inevitable) next question at some point?

[dpb]

I'm unaware of a specific list, no, altho I'm sure there are many around. These things that seem so obvious once pointed out are simply part of learning; almost everybody will have done the same or something very similar at some time in their evolution. I think it's probably a case where most of them one has to discover on their own simply as reinforcing the learning process--one can be told all these things but until experience for oneself it doesn't make the same impression.