The best (or worst?) counterexample to Schwarz theorem

Question

The well known theorem of Schwarz asserts the following: suppose that $f:U \to \mathbb{R}$ where $U \subset \mathbb{R}^n$ is $C^k$ function and pick some sequence $(j_1,...,j_k)$ of length $k$ where each $j_i \in \{1,2,...,n\}$. Then all mixed partial derivatives of $k$-th order over variables $x_{j_i}, i=1,2,...,k$ coincide, no matter in which order we diffrentiate. I would like to see the following example: function $f:\mathbb{R}^n \to \mathbb{R}$ which has all partial derivatives up to $k$-th order (where $k$ is a fixed positive integer) but for every point $x_0 \in \mathbb{R}^n$ all $k$-th order derivatives has different values. In other words I would like to see an example where for every permutation of variables we get different value of partial derivative and this happens not only in one but in every point.