Division-free determinant identity for column replacement
Proveddet_mul_det_updateColcommutative-ringscramers-ruledeterminantslinear-algebra
Let be a commutative ring, let be a finite index set, and let . Fix and a vector . Write for column replacement, and write for the th column of . Then
The identity requires neither an inverse of nor a nonzero determinant. It provides a division-free way to transfer Cramer's vector identity through a determinant that is linear in one column.
Preamble
import Mathlib.LinearAlgebra.Matrix.Adjugate open scoped BigOperators
Formal statement
theorem det_mul_det_updateCol
{ι R : Type*} [Fintype ι] [DecidableEq ι] [CommRing R]
(A B : Matrix ι ι R) (j : ι) (v : ι → R) :
A.det * (B.updateCol j v).det =
∑ i, (A.updateCol i v).det *
(B.updateCol j (fun r => A r i)).det := by sorrySource
An abstract consequence of Cramer's vector identity and determinant linearity. Mathlib at revision c5ea00351c28e24afc9f0f84379aa41082b1188f, LinearAlgebra/Matrix/Adjugate.lean, Matrix.mulVec_cramer (line 284) and Matrix.cramer / Matrix.cramer_apply (lines 94–98): https://github.com/leanprover-community/mathlib4/blob/c5ea00351c28e24afc9f0f84379aa41082b1188f/Mathlib/LinearAlgebra/Matrix/Adjugate.lean#L276-L285 . This is the division-free scalar identity derived from those results, not an additional claim about an invertible matrix.