vcat acts inconsistently when only some arguments are row vectors

[nicholaskl97]

If you mix row vectors and elements as arguments in vcat and all the row vectors are at the end, then it fills in the lone elements into full rows, but if any lone element comes after a row vector, then vcat throws a dimension mismatch error. Probably, vcat should either always throw the error or always fill in, regardless of argument order. I couldn't find anything in the documentation for vcat saying what behavior a user should expect in this case.

vcat("foo", ["bar" "hello" "world"]) outputs

2×3 Matrix{String}:
 "foo"  "foo"    "foo"
 "bar"  "hello"  "world"

but
vcat(["bar" "hello" "world"], "foo") throws an error:

ERROR: DimensionMismatch: mismatch in dimension 2 (expected 3 got 1)
Stacktrace:
  [1] _cs
    @ ./abstractarray.jl:1717 [inlined]
  [2] _cshp
    @ ./abstractarray.jl:1703 [inlined]
  [3] _cshp
    @ ./abstractarray.jl:1714 [inlined]
  [4] _cat_size_shape
    @ ./abstractarray.jl:1693 [inlined]
  [5] cat_size_shape(dims::Tuple{Bool}, X::Matrix{String}, tail::String)
    @ Base ./abstractarray.jl:1691
  [6] _cat_t(::Val{1}, ::Type{String}, ::Matrix{String}, ::Vararg{Any})
    @ Base ./abstractarray.jl:1732
  [7] _cat(::Val{1}, ::Matrix{String}, ::Vararg{Any})
    @ Base ./abstractarray.jl:1728
  [8] #cat#155
    @ ./abstractarray.jl:1916 [inlined]
  [9] vcat(::Matrix{String}, ::String)
    @ Base ./abstractarray.jl:1815
 [10] top-level scope
    @ REPL[15]:1

Above examples used strings, but the same is true with numbers:
vcat(1, 2, [3 4 5], [6 7 8]) outputs

4×3 Matrix{Int64}:
 1  1  1
 2  2  2
 3  4  5
 6  7  8

but vcat([3 4 5], [6 7 8], 9) throws an error:

ERROR: DimensionMismatch: mismatch in dimension 2 (expected 3 got 1)
Stacktrace:
 [1] _cs
   @ ./abstractarray.jl:1717 [inlined]
 [2] _cshp
   @ ./abstractarray.jl:1703 [inlined]
 [3] _cshp
   @ ./abstractarray.jl:1714 [inlined]
 [4] _cat_size_shape (repeats 2 times)
   @ ./abstractarray.jl:1693 [inlined]
 [5] cat_size_shape
   @ ./abstractarray.jl:1691 [inlined]
 [6] _cat_t
   @ ./abstractarray.jl:1732 [inlined]
 [7] typed_vcat(::Type{Int64}, ::Matrix{Int64}, ::Matrix{Int64}, ::Int64)
   @ Base ./abstractarray.jl:1880
 [8] vcat(::Matrix{Int64}, ::Matrix{Int64}, ::Int64)
   @ LinearAlgebra /opt/julia/julia-1.8.5/share/julia/stdlib/v1.8/LinearAlgebra/src/special.jl:419
 [9] top-level scope
   @ REPL[22]:1

versioninfo() outputs

Julia Version 1.8.5
Commit 17cfb8e65ea (2023-01-08 06:45 UTC)
Platform Info:
  OS: Linux (x86_64-linux-gnu)
  CPU: 20 × 12th Gen Intel(R) Core(TM) i9-12900H
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-13.0.1 (ORCJIT, goldmont)
  Threads: 20 on 20 virtual cores
Environment:
  JULIA_NUM_THREADS = auto

I installed Julia via the official generic binaries, as on julialang.org.

[oscardssmith]

wow, that's bad. These should all through.

[BioTurboNick]

I previously documented this. #38019