usually, symbolic functions return vectors for vector inputs:
syms('x');
f=x*2;
subs(f,[1 2 3])
outputs: [2 4 6]
but doing
f=sym('0');
subs(f,[1 2 3]);
outputs: 0
and not: [0 0 0]
so basically, my questions is how do I make f behave as a "normal" symbolic function.
I can do something ugly like f=x-x to create a function that always returns zero, but is there a prettier way?
Actually, there is no way.
sym('0') creates a symbolic constant (0, in this case). subs() replaces all variables with each value from the given vector. However, you have no variables, so subs() just returns the given symbolic constant.
It gets better. sym() internally does some simplification, so sym('0*x') or sym('x-x') both become sym('0') and you get the exact same behavior. Similarly, sym('x/x') turns into sym('1') and you just get the scalar 1 back from subs() even if you pass it a vector.
The only way you're going to get around this is by writing a helper function that detects if the size() of the output from subs() is less than the vector, and turns it into the correct size of a vector if needed.
sym('x-x') and it did produce the desired effect.
I don't have the Symbolic Toolbox so this is a blind guess. Your function declarations don't look identical enough:
syms('x');
f=x*2;
against
f=sym('0');
Perhaps what you have done in the first case is define a function which returns the double of its inputs, in the second case defined a function which returns 0 whatever its inputs. Maybe
syms('x');
f=x*0;
is what you need ?
