Hi. I am relatively new to Julia. Suppose I have a matrix A and I write v = A[:, 1]. Now v is a vector as expected. But u = A[1, :] is a vector again (I expected u to be adjoint instead of vector). What would the problems be u were an adjoint instead of being vector? I have gone through the thread…

This talk might be helpful: [JuliaCon 2017 | Taking Vector Transposes Seriously | Jiahao Chen]

Thank you for the link to the talk. It is a very nice talk and I had watched it earlier. I watched it again to see if I had missed something. The summary of the talk is that one requires four objects Scalars Column vector Row Vector Matrices to implement householder notation. Please correct me …

Thank you for pointing to the discussion. From comment of lucas711642 I understand that the reason for this behavior is that A[1,:] being an adjoint does not generalize to multidimensional arrays with dimension greater than 2.

Matrix operations not symmetric

Internals & Design

simeonschaub October 6, 2021, 5:42am 4

This has already been discussed multiple times in this forum. See Indexing into Matrix for example and the linked post within.

Topic		Replies	Views
Indexing into Matrix General Usage	7	5598	August 31, 2020
Vector adjoint and indexing don't commute? General Usage	19	1564	September 5, 2020
Vector vs Column vector New to Julia question	13	9744	February 21, 2020
Transposes vs 1-row matrices Internals & Design linearalgebra	3	440	April 8, 2024
Confusing Results When Comparing Arrays in Julia General Usage question	5	206	September 22, 2024

Matrix operations not symmetric

Related topics