I am stuck at this counting. I am explaining my question with an example.
Suppose we take $n=2\times 3\times 5\times 7$. I want to find those elements in the set $$X=\{7,14,21,28,35,42,49,56,\ldots ,203\}=\{7k\,|\, 1\leqslant 7k<n\}$$ which have $2,3,5$ as one of their prime factors and count their number. I call this number $x$.
The numbers in the set $X$ which have $2,3,5$ as one of their prime factors are $$\{14,21,28,35,42,56,63,70,84,98,105,112,126,140,147,154,175,182,189,196\}.$$
I took the subgroup generated by $2\times 7=14$ which is $\{0,14,28,42,\ldots,196\}$. I took the subgroup generated by $3 \times 7=21$ which is $\{0,21,42,63,\dots,189\}$. I took the subgroup generated by $5\times 7=35$ which is $\{0,35,70,105,\dots,175\}$.
I found $x=$ number of elements in subgroup generated by $14$ + number of elements in subgroup generated by $21$+number of elements in subgroup generated by $35$ -number of elements in subgroup generated by $42$-number of elements in subgroup generated by $70$-number of elements in subgroup generated by $105=15+10+6-6-3-2=20$.
My question is:
If $n=p_1^{\alpha_1}p_2^{\alpha_2}\cdots p_k^{\alpha_k}$ and if we consider the cyclic subgroup generated by $p_k$ denoted by $\langle p_k\rangle$, what will be the number of those elements in $\langle p_k\rangle$ which have $p_i$ where $1\le i\le k-1$ as one of their prime factors?
I have provided as much detail as I could. If I need to provide any other details please ask in the comment box
Regarding the example: Let's allow $0$ in $X$, then $X = 7\mathbb{Z}/n\mathbb{Z} \cong \mathbb{Z}/m\mathbb{Z} =: R$, where $m = 2\cdot 3\cdot 5$. You are searching for the number $x$ of Elements in $X$ which share a factor with $|R| = m$. Below I show that the number of $k\in\mathbb{N}, k\leq m$ which are coprime to $m$ is equal to the number of units (invertible elements) in $R$, which is given by $\varphi(m)$, where $\varphi$ is Euler's totient function. Finally you get $$x = m - \varphi(m) - 1 = 30 - 8 - 1 = 21 $$ (I'm subtracting $1$, because $X$ contains $0$ now. I don't know why our counts are off by one, though)
Edit: I just noticed I haven't really answered your generalized question, but I'd say it works just like in the special case from above.
Proof of the claim from above.
Let $\varphi : \mathbb{N} \longrightarrow \mathbb{N}, n \longmapsto |E(\mathbb{Z}_n)|$ and $M_n := \{k\in\mathbb{N} \mid k \leq n,\; \gcd(k,n) = 1 \}$. We show $|M_n| = \varphi(n)$. Choose an arbitrary $n\in\mathbb{N}$.
''$\leq$'': Let $m \in M_n$. Then the Euclidian algorithm gives $k,\ell \in \mathbb{Z}$ such that $1 = km + \ell n$. I.e., $\bar 1 = \overline{km} + \overline{\ell n} = \overline{km}$ in $\mathbb{Z}_n$, thus $(\bar m)^{-1} = \bar k$ and thus $\bar m \in E(Z_n)$. Because $m \leq n$ for all $m \in M_n$, we get $\bar m' \neq \bar m$ for $m'\neq m \in M_n$.
''$\geq$'': Let $\bar m \in E(\mathbb{Z}_n)$. Then there is $k \in \mathbb{Z}$ such that $1 = \overline{mk}$, hence there is $\ell \in \mathbb{Z}$ such that $1 = km + \ell n$. Thus $1 \in m\mathbb{Z} + n\mathbb{Z} = d\mathbb{Z}$. Such a $d\in\mathbb{Z}$ exists, because $\mathbb{Z}$ is a principle ideal domain. Then $d$ is a gcd of $n$ and $m$. From $1 \in d\mathbb{Z}$ follows $d\mathbb{Z} = \mathbb{Z}$ and therefore $d\in \{1,-1\}$.