Inv(transpose()) fails on 1x1 matrix

New to Julia
1

Hello, may have found a bug about inv() and transpose():

julia> mat = reshape([1.0 2.0; 3.0 4.0], 2, 2)
2×2 Array{Float64,2}:
 1.0  2.0
 3.0  4.0

julia> inv(mat)
2×2 Array{Float64,2}:
 -2.0   1.0
  1.5  -0.5

julia> inv(transpose(mat) )
2×2 Array{Float64,2}:
 -2.0   1.5
  1.0  -0.5

julia> mat1 = reshape([3.0], 1, 1)
1×1 Array{Float64,2}:
 3.0

julia> inv(mat1)
1×1 Array{Float64,2}:
 0.3333333333333333

julia> inv(copy(transpose(mat1) ) )
1×1 Array{Float64,2}:
 0.3333333333333333

julia> inv(transpose(mat1) )
ERROR: MethodError: no method matching ldiv!(::Float64, ::Array{Float64,2})
2

Either my basic math is failing miserably or that result is exactly correct…
(Edit: Sorry, forget it please, my browser was hiding the last command.)

3

I suggest editing your post to break out this last line. It is very easy to miss right now.

4

A fix would be to define:

LinearAlgebra.ldiv!(a::Number, B::AbstractVecOrMat) = ( B ./= a )
1 Like
5

The same issue is present for adjoint and @Per’s suggestion also fixes that.

6

This is related to (but not exactly the same as the first issue mentioned at) https://github.com/JuliaLang/julia/issues/28782