Hi, How we can create a column matrix (i.e. nx1) with array literals? I am aware of other ways for achieving this such as using initializer functions such as “zeros” and “ones”. but how we can do this with array literals? I expect the following snippet create a column matrix:
A = [1; 2]
and also following snippet:
A = [1
2]
but both of this will create a 1 dimensional array as “2-element Array{Int64,1}”.
in Documetation explicitly states:
Concatenation
If the arguments inside the square brackets are separated by semicolons ( ; ) or newlines >instead of commas, then their contents are vertically concatenated together instead of the >arguments being used as elements themselves.
I think this is a little inconsistency in Julia syntax.
1 Like
No, because 1d arrays are treated as column vectors in all of the Julia linear-algebra routines. (Because of this, it’s not clear why you need to create a 1-column matrix instead of a 1d array?)
You could use reshape([1,2],2,1).
(It might also make sense to have a constructor Matrix(x::Vector) = reshape(x, length(x), 1); currently there is no such constructor defined in Base.)
See also this discussion and issue #17084.
Thank you for your respond. my point is if semicolon or newline concatenate vertically why it does not work in this case?
Because it is doing “vertical” concatentation (vcat) — from Julia’s perspective, a 1d array is “vertical”, and more generally the first dimension of any array (of any dimensionality) is defined to be “vertical”. For the same reason:
julia> [[3,4]; [5,6]]
4-element Array{Int64,1}:
3
4
5
6
julia> size([rand(3,4,5); rand(17,4,5)])
(20, 4, 5)
1 Like
Thank you. your explanation helped.