Saving output of for loop after each iteration (for a function)

y = 0:0.05:1;
u= 2*y-(2*y.^3)+y.^4;
a = 0:0.01:1;
b = 4000; 

Carray=zeros(1,101)
for i=1:size(a)
        c=A_function(y,u,b,a(i));
        Carray(i)=c;
end

I want to run the function 'c' for all values of 'a'

and I want to store value of 'c' for each 'a'

The error i get is "Unable to perform assignment because the left and right sides have a different number of elements"

Can someone please let me know what I am doing wrong?

Thanks

Answer 1

[Sulaymon Eshkabilov]

Here is the corrected code of yours:

y = 0:0.05:1;
u= 2*y-(2*y.^3)+y.^4;
a = 0:0.01:1;
b = 4000; 
C=zeros(numel(a),numel(u));
for ii=1:size(a)
        C(ii,:)=A_function(y,u,b,a(ii));
end

[JD]

Now I get the error “unable to perform assignment because the size of the left side is 1-by-101 and the size of the right side is 100-by-1”

I tried C(ii-1,:) = A_function(y,u,b,a(ii)); But that was an invalid input.

Also I think that the array size row vs Columns is creating a problem left to right?

Thanks

[JD]

I fixed the error by changing this line “For ii=1:size(a)-1”

But now my C matrix is still all 0’s

[Steven Lord]

%{
C=zeros(numel(a),numel(u));
for ii=1:size(a)
%}

I agree with the first line but not the second. Before looking at the output of the code below, how many lines of text does it display?

a = 0:0.1:1;
for k = 1:size(a)
fprintf("k is %d\n", k)
end

k is 1

Now compare this with using numel instead of size.

a = 0:0.1:1;
for k = 1:numel(a)
fprintf("k is %d\n", k)
end

k is 1
k is 2
k is 3
k is 4
k is 5
k is 6
k is 7
k is 8
k is 9
k is 10
k is 11

Why the difference in behavior? When any of the inputs to the colon function or : operator is non-scalar and non-empty, MATLAB will just use the first element of that input.

sz = size(a)

sz = 1×2

1 11

v = 1:sz % equivalent to 1:sz(1)

v = 1

While your code would have worked for a column vector, the numel approach works in general to allow you to loop through all the elements in an array.

Answer 2

[Image Analyst]

Try this:

y = 0:0.05:1;
u = 2*y-(2*y.^3)+y.^4;
a = 0:0.01:1;
b = 4000;

rows = length(a)
columns = 101;
Carray = zeros(rows, columns);
for row = 1 : rows
	resultVector = A_function(y, u, b, a(row));
	Carray(row, :) = resultVector;
end

function c = A_function(y, u, b, aValue)
c = randi(9, 1, 101); % Whatever it may be...
end

Obviously replace the A_function with your actual function that returns a vector. Carray will not be all zeros.

[JD]

Thanks Image Analyst. I get the error “Unable to perform assignment because of the size of the left side is 1-by-101 and the size of the right side is 101-by-101”

This error is for line Carray(row, :) = resultVector;

I changed it to Carray(row, columns) = resultVector; But that made it a 1-by-1

Also what is ‘aValue’ in function c?

Thanks

[JD]

It appears that the error is in the Carray(row, :) = resultVector; Line

Because before I run that line row=101 But after I run it, row =1

[Image Analyst]

My code runs fine with no errors. If you have errors it's because you replaced the A_function with the actual function, and you did not supply that to us. Please attach it.

When you call A_function you pass it a(i) which is a single value, not a vector. That is fine, however, you cannot define the function like this:

function c = A_function(y, u, b, a(i))

because it will throw an error. The 4th argument needs to be a simple variable name that represents a(i) and I called it aValue.

[JD]

function c=OS_Eigen2D(y,u,b,a,N)

% Set up as a function, the following input parameters are required.

% INPUT:
% y  
% u 
% b 
% a
% N     =   (Optional) 

Discretization order for Chebyshev polynomials
%           N is defaulted to be N= 101 if otherwise specified; No need to
%           change in general

% OUTPUT
% c  

% N is the number of data points for the Gauss-Lobato interpolation (optional)
 if ~exist('N','var')
     % N parameter does not exist, so default it to something
      N = 101;
 end

%%SET-UP DIFFERENTIATION MATRICES
[D,yGL]=cheb(N);
D2=D^2;
D2=D2(2:N,2:N);
S=diag([0; 1./(1-yGL(2:N).^2); 0]);
D4=(diag(1-yGL.^2)*D^4-8*diag(yGL)*D^3-12*D^2)*S;
D4=D4(2:N,2:N);
D0=eye(N-1);

%%INTERPOLATE TO GAUSS-LOBATTO POINTS AND TRANSFORM COORDINATES
ymax=max(y);
y_trnsfrm=2.*y./ymax-1;
dzdy=2/ymax;

uGL=interp1(y_trnsfrm,u,yGL,'spline');
dduGL=(dzdy^2)*(D^2)*uGL;

uGL=uGL(2:end-1);
dduGL=dduGL(2:end-1);

D2=(dzdy^2)*D2;
D4=(dzdy^4)*D4;

%%DISCRETIZE O-S EQUATION
ak2=a^2;

R0=D2-ak2*D0;
R1=(1i/b/a)*(D4-2*ak2*D2+(ak2^2)*D0);
R1=R1+(uGL*ones(1,length(uGL))).*R0-(dduGL*ones(1,length(uGL))).*D0;

%%SOLVE
c=eig(R1,R0);

end

%%Chebyshev  compute D = differentiation matrix, x = Chebyshev grid
 function [D,x] = cheb(N)
  if N==0, D=0; x=1; return, end
  x = cos(pi*(0:N)/N)'; 
  c = [2; ones(N-1,1); 2].*(-1).^(0:N)';
  X = repmat(x,1,N+1);
  dX = X-X';                  
  D  = (c*(1./c)')./(dX+(eye(N+1)));      % off-diagonal entries
  D  = D - diag(sum(D'));                 % diagonal entries
 end

[Image Analyst]

Unrecognized function or variable 'alp'.

Error in test7>OS_Eigen2D (line 74)

ak2=alp^2;

Error in test7 (line 20)

resultVector = OS_Eigen2D(y, u, b, a(row));

What is alp?

[JD]

Sorry alp should be ‘a’

Forgot to change that before posting. I will update original function code as well

[JD]

Hi Image Analyst,

Did that work for you? I have to turn in my work tonight so just wondering if you were able to troubleshoot why it’s not working with the actual function.

Thanks

[Image Analyst]

Sorry, I didn't get a chance last night, but I guess it's past the homework deadline so it doesn't matter at this point.