There are several ways to argue that the $q$-adic valuation on $\mathbb Q$ can be extended to a valuation $\mathsf v_q$ on the field $\mathbb R$ of real numbers.
I know that such a valuation cannot be explicitely given. Nevertheless I suspect that it is possible to compute very special values.
I summarize two questions here, because they are related. Denote by $\mathbb P$ the set of prime numbers.
- Let $(a_p)_{p \in \mathbb P}$ be a sequence of rational numbers such that $\prod_{p \in \mathbb P} a_p$ is convergent (wrt euclidean topology). Let $q \in \mathbb P$ and assume that $\sum_{p \in \mathbb P} \mathsf v_q(a_p)$ is also convergent. Does it follow that $\mathsf v_q(\prod_{p \in \mathbb P} a_p) = \sum_{p \in \mathbb P} \mathsf v_q(a_p)$?
An obvious proof attempt is to show that $\mathsf v_q(\prod_{p >n} a_p)$ tends to $0$ as $n$ goes to infinity. This would of course be easy if $\mathsf v_p$ would be continuous with respect to the euclidean topology.
- The understanding of 1. could lead to constructions of real numbers with prescribed $q$-adic valuations. For instance, does there exists $x \in \mathbb R$ such that $\mathsf v_q(x) = 1$ for all $q \in \mathbb P$?
Thank you in advance for your help!
$\sum_{p}(v_{q}(a_{p}))$ is given to be convergent. Now, as the discrete valuation takes discrete values, preferably maybe from $Z$, the above series is convergent $\iff$ it is eventually zero. Which means question 1 boils down to finite case for the same problem, to which the answer is yes. Which I believe leads to a vacuous answer for question $2$.