Consider computable functions $f: \mathbb{N} \rightarrow \mathbb{N}$, given as formulas. I assume that it is clear for at least some of them whether they contain a distinction of cases:
$$f(n) = n$$
obviously does not contain a distinction of cases (i.e. is d.c.-free),
$$ g(n) = \begin{cases} 1 & \text{for } n = 0 \\ n & \text{otherwise} \end{cases} $$
obviously does, at least at the face of it.
What is not clear is whether there is a d.c.-free function $g'$ that is equivalent with $g$, is it?
What I believe to know:
It is not decidable whether there is a d.c.-free function equivalent with a given one (unless one can prove that there is one for every computable function!)
Especially, a d.c.-free function $g'$ is not "computable" (as a formula) from a given one $g$
Even when you are given one (by an oracle): showing the equivalence of $g'$ and $g$ is not decidable.
Nevertheless there are at least some trivial cases where a d.c.-free formula exists, e.g. for
$$ g(n) = \begin{cases} 0 & \text{for } n = 0 \\ n & \text{otherwise} \end{cases} $$
Can anyone provide a non-trivial example?