Evaluating a cell array of function handles

In the following code I assemble a sequence of polynomials (the Chebyshev polynomials, to be exact) symbolically, and then use matlabFunction() to convert each symbolic expression into a function handle. Each function handle is then placed in the cell array.

N = 10;
C = cell(N,1);

syms y;
T = sym(zeros(N,1));

T(1) = 1;
T(2) = y;

C{1} = matlabFunction(T(1));
C{2} = matlabFunction(T(2));

for n = 2:1:N-1
    T(n+1) = 2*y*T(n) - T(n-1);
    C{n+1} = matlabFunction(T(n+1));
end

evalArray = @(y)cellfun(@(f)f(y),C);

evalArray(3)

However, evalArray(3), which is intended to evaluate each polynomial at y=3, gives the error: "Error using symengine>@()1.0;Too many input arguments."

Any suggestions? Thanks all for your help.

Answer 1

[Matt Fig]

Exactly as you have made it, C{1} is a function that takes no input arguments.

Try this instead:

C{1} = matlabFunction(T(1),'vars',y);

[Walter Roberson]

Right. By default matlabFunction searches for variables in the symbolic code and uses that many variables as input; since "1" is constant (0 variables), that is 0 inputs; Matt's solution shows how to force one input.

[Matt Fig]

matlabFunction is even smarter than that!

T(1) = 1 + 0*y;  % Fails to produce f = f(y)
T(1) = y - 2 + 3 - y;  % Also fails!

I tried these and other ways of "tricking" the conversion, without success. Finally, I checked the doc for matlabFunction... d'oh.