Unexpected behaviours of spdiagm

[phipsgabler]

Say we have

julia> y = sprand(10, 0.2)
10-element SparseVector{Float64,Int64} with 2 stored entries:
  [2 ]  =  0.804026
  [3 ]  =  0.0159904

Then, firstly, there is no method that works implicitly on the main diagonal, like diagm(y) does -- you have to explicitely say spdiagm(0 => y). Secondly, and more importantly, this method does not preserve sparsity of the input:

julia> spdiagm(0 => y)
10×10 SparseMatrixCSC{Float64,Int64} with 10 stored entries:
  [1 ,  1]  =  0.0
  [2 ,  2]  =  0.804026
  [3 ,  3]  =  0.0159904
  [4 ,  4]  =  0.0
  [5 ,  5]  =  0.0
  [6 ,  6]  =  0.0
  [7 ,  7]  =  0.0
  [8 ,  8]  =  0.0
  [9 ,  9]  =  0.0
  [10, 10]  =  0.0

Where I really had expected the result to be

julia> spdiagm(0 => y)
10×10 SparseMatrixCSC{Float64,Int64} with 2 stored entries:
  [2 ,  2]  =  0.804026
  [3 ,  3]  =  0.0159904

(That's on 1.3.1, sorry for not having updated yet :))