Let $\mathbb F_5=\{0,1,2,3,4\}$ is finite field of size $5$. I am trying to find minimal $n$ so that vector space of dimension $n$ over $\mathbb F_5$ contains $2$ linearly independent vectors so that all vectors of their linear span contains $1$ as a coordinate. In other words there should be linear subspace of dimension $2$ so that all vectors contains $1$ coordinate (except trivial vector).
For $n=8$ there is a solution: $(1,2,3,4,0,0,0,0)$ and $(0,0,0,0,1,2,3,4)$.
I need hint to prove that no solution for $n<8$ exist.
This is a tough nut to crack (or, more likely, I'm looking at it all wrong). Anyway, I think I found a promising way to look at it.
Conjecture. It is impossible to cover all the points of the punctured plane $\Bbb{F}_5^2\setminus\{(0,0)\}$ by seven lines, none passing through the origin.
This is just a conjecture. I think it's true, but I'm not betting a fortune on it yet. At the bottom I will share my thoughts, but let's first show how this settles your question.
Claim. There aren't any 2-dimensional subspace of the prescribed type for $n<8$.
Proof. (Assuming the above conjecture.) Assume contrariwise that such a subspace $V$ exists. Let $u=(u_1,u_2,\ldots,u_n)$ and $v=(v_1,v_2,\ldots,v_n)$ be a basis of $V$. The required property says that for all $(x,y)\in \Bbb{F}_5^2\setminus\{(0,0)\}$ there exists an index $i, 1\le i\le n$ such that $xu_i+yv_i=1$. This implies that the set $\Bbb{F}_5^2\setminus\{(0,0)\}$ is covered by the $n$ lines $u_ix+v_iy=1$. Clearly none of those pass through the origin. Therefore the conjecture implies that $n>7$.
Remark. Observe that 2-dimensional subspace of $\Bbb{F}_5^6$ spanned by $(0,1,1,1,1,1)$ and $(1,0,1,2,3,4)$ has the property that all the non-trivial vectors have exactly one component equal to zero. Getting a fixed non-zero component to appear is more difficult as witnessed here. Of course, scalar multiplication allows us to replace $1$ with any other non-zero scalar.
How might the conjecture be verified? The lines in the plane $\Bbb{F}_5^2$ can have one of six directions, either parallel to the $y$-axis or one of the five possible slopes. Therefore among the seven lines (assuming that the conjecture could be false) there is at least one parallel pair. This leads to a systematic (IMO manageable) case-by-case analysis.