problem with symbolic to function handle

Hi, i have a symbolic function of 3 variables and then I compute its gradient and convert each entry of the gradient to a function handle,but some of them become function handles of less than 3 variables, and that makes the evaluation of the vector field more difficult, any ideas on how to solve this?

syms x y z 

f=x*y*z+x*2*y

G=gradient(f)

Dx=matlabFunction(G(1));

Dy=matlabFunction(G(2));

Dz=matlabFunction(G(3));

[X,Y,Z]=meshgrid(-3:1:3);

Fx=Dx(Y,Z);

Fy=Dy(X,Z);

Fz=Dz(X,Y);

this code works but i would like it to work when i write:

Fx=Dx( X,Y,Z);

Fy=Dy(X,Y,Z);

Fz=Dz(X,Y,Z);

and this is what my function handles look like:

Dx =

@(y, z) y .* z + 2 .* y

Dy =

@(x, z) x .* z + 2 .* x

Dz =

@(x, y) x .* y

Answer 1

[Steven Lord]

You can specify the list of input arguments the generated function or function handle should have using the 'Vars' parameter described in the documentation for the matlabFunction function. See the "Specify Input Arguments for Generated Function" example on that documentation page to see how to use this parameter.

[lorenzo donadio]

I dont underestand the answer in that link and how to use it, could you it to explain me?

[lorenzo donadio]

I found it thank you :D

[Walter Roberson]

m = matlabFunction(G, 'Vars', {x, y, z});

Answer 2

[Mischa Kim]

How about

   syms x y z

f = x*y*z+x*2*y;
G = gradient(f);
m = matlabFunction(G);

[X,Y,Z] = meshgrid(-3:1:3);
m_eval  = m(X,Y,Z);

[lorenzo donadio]

this works but it doesnt solve the problem, because if i change the f function to lets say x^2+y^2+z now the gradient is only a function of x and y and doing

m = matlabFunction(G);

and then m_eval = m(X,Y,Z);

gives me and error,how can i fix it?