declaration of global variables

Hello, I think I have declared global variable correctly but it keeps saying "Error: File: Ex3_6.m Line: 10 Column: 16 The GLOBAL or PERSISTENT declaration of "n" appears in a nested function, but should be in the outermost function where it is used." I declared it in the outermost function.

Can you guys help me?

==============================================

function [avg,med]=Ex3_6(in)

global n

n=length(in);

avg=newmean(in);

med=newmedian(in);

      function a=newmean(in)
          global n
          a=sum(in)/n;
      end

      function m=newmedian(v)
          global n
          w=sort(v);
          if rem(n,2)==1
              m=w((n+1)/2);
          else
              m=(w(n/2)+w(n/2+1))/2;
          end
      end

end

=================================

And I run below.

=====================================

score= score=[87 75 98 100 45 37 73];

[avg,med]=Ex3_6(score)

Answer 1

[the cyclist]

The proper syntax is to define the subfunctions outside of the "function ... end" construct:

function [avg,med]=Ex3_6(in)
    global n
    n=length(in);
    avg=newmean(in);
    med=newmedian(in);
end

function a=newmean(in)
  global n
  a=sum(in)/n;
end

function m=newmedian(v)
  global n
  w=sort(v);
  if rem(n,2)==1
      m=w((n+1)/2);
  else
      m=(w(n/2)+w(n/2+1))/2;
  end
end

[Jinho Jung]

I see. Thank you so much!

Another question is "can I define subfunctions inside the outermost function in any case like I did for the first time?"

I have run the function below and it works. This is why I get confused.

function z = funEx

for index=1:3

    tmp=fun2;
end
z=tmp;

    function y=fun2
        persistent x

        if isempty(x)
            x=0;
        end
        x=x+1;
        y=x;

    end
end

[the cyclist]

Actually, my brain must have been a bit switched off for the night, because I wasn't really thinking about your nested functions properly. (I almost never use nested functions.)

That being said, please do follow John's advice in his answer. (I was going to follow up with similar advice, but he beat me to it!) Not only should one avoid global variables in general, but nested functions specifically keep the parent function's variables within scope, by design.

Answer 2

[John D'Errico]

I'm sorry, but this is just about the silliest reason to use a global variable I've ever seen.

n is declared as global because you don't want to pass in n as an argument to the function, perhaps because you think this is more efficient?

Worse, these are nested functions. You never needed to pass in n at all, or make it global, since they can see the variables in the workspace of the caller.

A common result of globals is buggy code. What do you have? Buggy code. Surprise.

Were you told you had to use global variables as part of this assignment? I hope not, since then your teacher would be teaching you tools you should be avoiding in the first place. Far better they should be showing you how to pass data around properly.

function [avg,med]=Ex3_6(in)
n=length(in);
avg=newmean(in);
med=newmedian(in);

    function a=newmean(in)
      a=sum(in)/n;
    end

    function m=newmedian(v)
      w=sort(v);
      if rem(n,2)==1
        m=w((n+1)/2);
      else
        m=(w(n/2)+w(n/2+1))/2;
      end
    end
  end

As you can see above, I never defined n as global, and since those are nested functions, they can see n.

Or, you could have made them subfunctions. In that case, pass in n as an argument, or as good, it hurts nothing if you just obtain the value of n in each subfunction from the vector passed in.

[Stephen23]

+1 correct and entertaining, as always.