What is the Julian equivalent of Python’s np.eye(5, 5, 1) % python -q >>> import numpy as np >>> np.eye(5, 5, 1) array([[0., 1., 0., 0., 0.], [0., 0., 1., 0., 0.], [0., 0., 0., 1., 0.], [0., 0., 0., 0., 1.], [0., 0., 0., 0., 0.]])

Probably using diagm() . In your case: diagm(1 => [1, 1, 1, 1]); If you want efficient data structure you may use BandedMatrices.jl . I am not aware of a data structure specialized for the case the diagonals have the values 0 or 1 (Namely no need for multiplication for Matrix Multiplication).

It’s I in the LinearAlgebra standard library: In [1]: using LinearAlgebra In [2]: I(5) 5×5 Diagonal{Bool, Vector{Bool}}: 1 ⋅ ⋅ ⋅ ⋅ ⋅ 1 ⋅ ⋅ ⋅ ⋅ ⋅ 1 ⋅ ⋅ ⋅ ⋅ ⋅ 1 ⋅ ⋅ ⋅ ⋅ ⋅ 1

Note that the third index shifts the diagonal up creating an upper shift matrix instead of an identity.

Almost diagm(1 => [1, 1, 1, 1]) because of 5x5. Thanks!

Ah, apologies - an unusual use of eye then, given that the name as I understand it comes from Matlab and is a pun on the Identity matrix.

What is the Julian equivalent of Python's `np.eye(5, 5, 1)`

New to Julia

Topic		Replies	Views
Linear Algebra tricks? General Usage	28	2225	May 4, 2020
Eye in julia 0.7 New to Julia	36	8356	March 23, 2018
Transposed matrix New to Julia linearalgebra	1	475	February 17, 2022
Spdiag Matlab function in Julia General Usage	7	4097	February 19, 2021
"Diagonal coordinate system" for Julia arrays New to Julia question	15	749	June 11, 2024

What is the Julian equivalent of Python's `np.eye(5, 5, 1)`

Related topics