using anonymous functions in nested functions returns error if function is constant and you want do build the derivate, why and how to handle?

So i have this kind of nested function system using math toolbox

function difftest
x=sym('x')
start
function start
  input='0' %input=0 or other number -> error
  y=str2func(['@(x)' (input)]) 
  z(x)=diff(y(x)) %error
  p(x)=diff(z(x))
  h(x)=diff(p(x))
end
end

and i need the first,second and third derivate of the function called from user input, when the user input='0' % or other constant value the anonymous function is correctly y(x)=0 but the derivate with diff(y(x)) returns an error! This is not the case if y(x)=x and all the derivates are correctly displayed. What am i doing wrong?

Thanks in advance, Phil

Answer 1

[Walter Roberson]

What you are doing wrong is failing to supply the name of the variable to differentiate with respect to. When you do not supply a variable then symbolic diff() needs to use symvar() to find a variable to differentiate with respect to, and when your expression does not name any variables that is going to fail.