Let $n$ is a non-negative integer,$f$ diferentiable $n$ times in a neighbourhood of $x_{0}\in \mathbf{R},(x_{0}-\delta ,x_{0}+\delta ),$ and $f^{(n)}(x),$ the $n$th derivative of $f$ ,is continuous at $x=x_{0}.$ Is there a real number $\delta_{1},0<\delta_{1}<\delta,$ such that $f^{(n)}(x)$ is continous over $(x_{0}-\delta_{1} ,x_{0}+\delta_{1} )?$
When $n=0,$it is not difficulty to find a counterexample:
Let $$x_{0}=0,\qquad f(x)=\begin{cases} x^{2}& \text{if}\quad x \text { is a irratioanl number},\\ 0& \text{if}\quad x \text { is a rational number}. \end{cases} \quad $$
Go a step further,I need to find some counterexamples for higher orders, but I'm stumped for that .Can anyone give me any hints on how to start it? Any help will be appreciated.
The answer to the title question is "no", as may be deduced by (repeatedly) integrating the function constructed in the answer to If $f'$ is differentiable at $a$ then $f'$ is continuous at $(a - \delta,a + \delta)$.
Your proposed function is arguably not a counterexample, since $f'$ exists at only one point.
Incidentally, the term "fixed point" in the title has a technical meaning ("an $x$ such that $f(x) = x$"); since that appears not to be your intent, you might consider rewording the title. :)