Let $P=\{x:\;a_i^Tx\leq b_i,\;\forall i\}$ be a Polyhedron in $\mathbb{R}^n$. Let $x$ and $y$ be two distinct basic feasible solutions.
Recall $I_x =\{i:\;a_i^Tx\leq b_i \}$
Suppose that rank $\{ a_i: i \in I_x ∩ I_y \} = n − 1$.
Then the line segment $L_{x,y} = \{ \lambda x + (1 − \lambda)y :\, \lambda \in [0, 1] \}$ is an edge of P. Moreover, if P is a polytope, then every edge arises in this way.
I found this definition online and I have a problem understanding parts of it. If P is defined the way it is, then the inequality holds for all $i$ and for all (basic) feasible solutions. Wouldn't that then mean that $I_x=I_y=I_x\cap I_y$? Because otherwise there are $i$ for which the inequality doesn't hold and therefore the solution isn't an element of the Polyhedron, which is a contradiction. This part is quite confusing to me and I would appreciate anyone helping to clear up the confusion.