A matrix of rank has a nonzero minor
ProvedLatticePrepro.exists_det_submatrix_ne_zeroLet be a field, let and be finite index sets, and let be a matrix of rank . Then has a nonsingular submatrix: there are row indices and column indices such that
This is one direction of the characterization of the rank as the size of a largest nonvanishing minor. It is used in Lemma A.6, the rank bound for the polynomial matrix of an evaluation table: a nonsingular minor of the evaluation at the generator of the moment field selects a square submatrix of whose determinant is a nonzero polynomial.
Formalization Note: The row and column choices are maps and (not required to be injective in the statement), the submatrix is A.submatrix r c, and the rank is Mathlib's Matrix.rank. For the statement is trivial, since the empty determinant is .
import Mathlib
open Polynomial Matrix
variable {K : Type*} [Field K] {m n : Type*} [Fintype n]theorem LatticePrepro.exists_det_submatrix_ne_zero [Fintype m] (A : Matrix m n K) {k : ℕ} (hk : A.rank = k) :
∃ (r : Fin k → m) (c : Fin k → n), (A.submatrix r c).det ≠ 0 := by sorry