Output of a function as input of another function

Suppose, I have a function which gives me two matrices A and B as output and I want to use these 2 matrices as inputs for another function, how do I do it? I don't want to copy paste the output matrices or type the entire matrices as input for the second function.

Answer 1

[Matt Fig]

Say your first function is named FUNC1, and your second function is named FUNC2. You do not specify if FUNC2 returns any value...

[A,B] = func1(args);
func2(A,B);

A real life example that you can copy/paste!

[I,J] = max(magic(5))  % Call the MAX function with two returns.
power(I,J)  % Pass return args from MAX to POWER. (I.^J)

Answer 2

[Matt J]

If you mean you'd like to call func2(func1) in a single statement, then don't return A and B as separate output arguments from your first function. Have your first function pack them in a cell array. I.e., instead of

function [A,B]=func1(n)

    A=rand(n);B=rand(n,2);

function C=func2(A,B)

    C=A*B;

you could instead do this

function AB=func1(n)

    AB={rand(n); rand(n,2)};

function C=func2(AB)

    C=AB{1}*AB{2};

And now you can do

    C=func2(func1(n));

[Peanut]

This is exactly what I want to do, but in my case func2 is unchangeable. This is still possible if I use an intermediate variable i.e.

function AB = func1(n)
    AB = {rand(n); rand(n)};
end

function C = func2(A, B)
    C = A+B
end

ABoutput = func1(1)
Coutput = func2(ABoutput{:})

However, I'd like to skip the intermediate line as it improves readability, something like this

Coutput = func2(func1(1){:})

Is there anyway the function output can be directly unpacked rather than having the intermediate variable?

[Matt J]

function AB = func1(n)
    AB = {rand(n); rand(n)};
end

function C = func2(AB)
    C = A{1}+B{1}
end

Coutput = func2(func1(1))

[Peanut]

As I mentioned before, I can't change func2 (as it's a matlab function). I'll clarify exactly what the situation is:

syms x y z
eq1 = x == y^2
eq2 = y == z-1
eq2C = num2cell(children(eq2))
subs(eq1,eq2C{:})

Ideally, eq2C is never created. Something like below:

syms x y z
eq1 = x == y^2
eq2 = y == z-1
subs(eq1,num2cell(children(eq2)){:})

[Matt J]

You're viewing subs as playing the role of func2 here. I can see why you can't modify subs directly, but I don't see why you couldn't write a wrapper for subs that will do the expansion for you:

function out=mysubs(eq,argCell)

 out=subs(eq,argCell{:});

end

and apply it to your example as follows

syms x y z
eq1 = x == y^2
eq2 = y == z-1
mysubs(eq1,num2cell(children(eq2)))

To make something generally applicable beyond just subs(), you could write a general utility function called, say, myfeval

function varargout=myfeval(fun,varargin)

  [varargout{1:nargout}]=feval(fun,varargin{1:end-1},varargin{end}{:});

end

and apply it to your example as,

syms x y z
eq1 = x == y^2
eq2 = y == z-1
myfeval(@subs,eq1,num2cell(children(eq2)))