hcat between sparse matrix and sparse vector

[blegat]

I would have imagined that concatenating a sparse matrix and a sparse vector would yield a sparse matrix but apparently this is not the case

Version 0.5.0-dev+4154 (2016-05-18 11:17 UTC)
Commit 778740e (11 days old master)
julia> [spzeros(2,2) spzeros(2)]
2×3 Array{Float64,2}:
 0.0  0.0  0.0
 0.0  0.0  0.0
julia> methods(hcat)
hcat(X::SparseMatrixCSC...) at sparse/sparsematrix.jl:2857
hcat{Tv,Ti}(X::AbstractSparseArray{Tv,Ti,1}...) at sparse/sparsevector.jl:704

It seems that there is no specific method so the general hcat is called. Is that on purpose or is there a definition missing ?

[tkelman]

The latter I think. A missed case from #15172.

[blegat]

Note that #15172 was closed 10 days ago and that nightly ubuntu's build are 11 days old, can you reproduce it on master ?

[zhmz90]

It still exists on master.

julia> [spzeros(2,2) spzeros(2)]
2×3 Array{Float64,2}:
 0.0  0.0  0.0
 0.0  0.0  0.0

julia> versioninfo()
Julia Version 0.5.0-dev+4393
Commit 15626c3* (2016-05-29 22:12 UTC)
Platform Info:
  System: Linux (x86_64-linux-gnu)
  CPU: Intel(R) Xeon(R) CPU E5-2699 v3 @ 2.30GHz
  WORD_SIZE: 64
  BLAS: libopenblas (USE64BITINT DYNAMIC_ARCH NO_AFFINITY Haswell)
  LAPACK: libopenblas64_
  LIBM: libopenlibm
  LLVM: libLLVM-3.7.1 (ORCJIT, haswell)

[pkofod]

Hm, as far as I remember, it was kind of on purpose. When I started working on the pr, I don't think sparse vec cat worked at all, I think it's in the discussion somewhere. I can make a pr to fix it.

edit found my comment #15172 (comment)

"I havn't done anything about the fact that [spvector spmatrix] doesn't exist, as it should probably be a different PR."

I'm on it.

edit2 I have a fix, but My computer is offline, and my phones internet sharing is broken. Will make a pr asap.