Cant create a function that return array correctly

Hello,

I'm learning how to use Matlab for some computational physics simulations and i'm having the following problem.

I define a certain quantity

x = [0:dx:L];
 y(1,:) = cos(x).*exp(-(x-1*L/2).^2/(2*20));

where the first index of the array represents time and the second (x) represents space.

I want to advance this quantity in time with a simple relation lets suppose, so i create a loop

 for n = 1:Nmax 

      for i = 1:Imax

            x(n+1,i) = x(n,i) - 0.5;
      end
 end

this work just fine..

But now imagine if i want to create a function that advances this quantity x in another .m file and i want to call it here as so

 for n = 1:Nmax 

            x = advancex(n,Imax,x);

 end

where the advance function (independent .m file)

function z =advancex(n,Imax,x)

      for i = 1:Imax
          z(n+1,i) = x-0.5;

      end

This way work poorly because only final array is zero in all position in time except for the final position i.e. only z(Nmax,:) is correctly calculated.

How do I do this breakdown correctly in such a way as to end up with an array which has the values of z at every time step n in all positions?

Open to suggestions, thank you so much :)

Answer 1

[dpb]

function z =advancex(n,Imax,x)
  for i = 1:Imax
    z(n+1,i) = x-0.5;

...only z(Nmax,:) is correctly calculated.

You don't have any thing except one n defined here in the function.

But, you don't really need a function or loops at all --

for n = 1:Nmax 
  for i = 1:Imax
    x(n+1,i) = x(n,i) - 0.5;
  end
end

Can be written as simply

 x(2:Nmax+1,:)=x(1:Nmax,:)-0.5;

NB that x(1,:) is unchanged from original here and the final size(x,1) is one row larger than the original.

[Pedro]

"You don't have any thing except one n defined here in the function."

Within the function your are correct, but since the function is in a loop i'm expecting it to work for every n... Do you agree?

"Can be written as simply

 x(2:Nmax+1,:)=x(1:Nmax,:)-0.5;
NB that x(1,:) is unchanged from original here and the final size(x,1) is one row larger than the original."

Great idea, i'll try this. It's just that in my actually program I have to advance these quantities with complicated algorithms and so im trying to create parallel functions to this this to avoid a main code that it to confusing.

Thanks for your help!

[dpb]

It does; but--somewhere you have to allocate x to contain more than the one row n which your code doesn't show how/where you initialize it. When you write

x(n,:) = something;

where x isn't allocated or size(x,1)<n, then Matlab allocates an array of size(n,RHS) or reallocates (extends) to that size. But, any elements in x that have not been preassigned will be zero. I would expect that there is where the rest of the zeros are coming from.

Again I caution you that creating arrays dynamically this way is EXTREMELY expensive in run time performance and is NOT recommended.