I'm working in the following exercise.
First, we define the outer Lebesgue measure, $m(A)=\inf\{\sum_n l(I_n): (I_n)\text{ is a sequence of open intervals and} A \subset \bigcup_nI_n \}$, where $l$ denotes the lenght of the interval.
Suppose $A \subset \mathbb R$ and suppose $m(A)<\infty$. Show that for every $\epsilon>0$ there exists a bounded set $A_1 \subset A$ such that $m(A_1^c\cap A)<\epsilon$.
Progress: I was thinking about defining the decreascing sequence of sets $J_n=(-\infty, -n) \cup (n , \infty)$ and showing that one of them must have $\mu(A \cap J_n)<\epsilon$, then the remaining would follow easily, but I'm not managing to do it.
If $A$ is measurable Let $\epsilon>0$
Let $A = \bigcup_{k\in \mathbb{N}} A_k$, where $A_k$ is bounded. Let $B_i = \bigcup_{k=1}^i A_k$ and $B^*_i = B_i \setminus B_{i-1}$ (with $B^*_1=B_1=A_1$). We readily see that $\mu(A) = \sum_{k=1}^\infty \mu(B^*_k)$
Since $\mu(A)<\infty$, $\sum_{k=1}^\infty \mu(B^*_k) < \infty$ and then there exists $N$ such that $\sum_{k=N}^\infty \mu(B^*_k) < \epsilon$.
But then $B_N$ is bounded and $\mu((B_N)^C \cap A) = \mu(A \setminus B_N) = \mu( \bigcup_{k=N}^\infty B^*_k) < \epsilon$
If $A$ is not measurable, let $M \supset A$ such that $M$ is measurable. There is a bounded set $B\subset M$ such that $m(M\setminus B) <\epsilon$ but then $m(A\setminus (B\cap A)) = m(A\setminus B) \leq m(M\setminus B) <\epsilon$ and $B\cap A \subset A$ is bounded.