Using TypedTables limited column number

[Thomas008]

Using TypedTables I construct a table. If the table has more than 31 columns, then trying to access an element of rows(t) seems to cause compile errors.
The following code runs in the julia repl, bug compiling with juliac results in error messages.

using TypedTables

const COLUMN_NAMES = (
    :c1,:c2,:c3,:c4,:c5,:c6,:c7,:c8,:c9,:c10,:c11,:c12,:c13,:c14,:c15,:c16,
    :c17,:c18,:c19,:c20,:c21,:c22,:c23,:c24,:c25,:c26,:c27,:c28,:c29,:c30,
    :c31,:c32
    )

const TUPLE_TYPE = Tuple{
    Vector{String},Vector{String},Vector{String},Vector{String},
    Vector{String},Vector{String},Vector{String},Vector{String},
    Vector{String},Vector{String},Vector{String},Vector{String},
    Vector{String},Vector{String},Vector{String},Vector{String},
    Vector{String},Vector{String},Vector{String},Vector{String},
    Vector{String},Vector{String},Vector{String},Vector{String},
    Vector{String},Vector{String},Vector{String},Vector{String},
    Vector{String},Vector{String},Vector{String},Vector{String}
    }

const VALUES = (
    ["1"],["2"],["3"],["4"],["5"],["6"],["7"],["8"],["9"],["10"],
    ["11"],["12"],["13"],["14"],["15"],["16"],["17"],["18"],["19"],["20"],
    ["21"],["22"],["23"],["24"],["25"],["26"],["27"],["28"],["29"],["30"],
    ["31"],["32"]
    )

function (@main)(args::Vector{String})::Cint
  nt = NamedTuple{COLUMN_NAMES, TUPLE_TYPE}(VALUES)
  t = Table(nt)
  r1 = rows(t)[1]
  return 0
end

Compiling results in 4 error messages.

Verifier error #1: unresolved call from statement Core._apply_iterate(Base.iterate, Core.tuple, %new()::Vector{String})::Tuple{Vararg{String}}
Stacktrace:
 [1] _totuple(T::Type{NTuple{32, String}}, itr::Base.Generator{@NamedTuple{a::Vector{String}, b::Vector{String}, ...

Verifier error #2: unresolved call from statement (NT::Type{@NamedTuple{a::String, b::String, ....

The 3rd and 4th error messages are similar.

[gbaraldi]

You are hitting an annoying thing where Tuples are type stable up to 32. Then they become type unstable

[Thomas008]

Is there work to increase the type-stability of tuples?
Is there even a constant in the julia compiler to set the maximal number of elements in a tuple such that it is type-stable?

[Thomas008]

There are further limitations when using TypedTables:

1. Filtering rows: causes compiler errors, if the number of columns is greater than 5:

using TypedTables

  function (@main)(args::Vector{String})::Cint
    t = Table(c1 = [1], c2 = [1], c3 = [1], c4 = [1], c5 = [1], c6 = [1])
    t1 = filter(r -> r.c1 == 1, t)
    return 0
  end

Verifier error #1: unresolved call from statement (f::Function)(Core.getfield(nt::@NamedTuple{c1::Vector{String}, c2::Vector{Int64}, c3::Vector{Int64}, c4::Vector{Int64}, c5::Vector{Int64}, c6::Vector{Int64}}, 1)::Vector{String})::Any
Stacktrace:
 [1] map(f::Function, t::Tuple{Vector{String}, Vararg{Vector{Int64}, 5}})
   @ Base tuple.jl:357 [inlined]
 [2] map(::Function, ::@NamedTuple{c1::Vector{String}, c2::Vector{Int64}, c3::Vector{Int64}, c4::Vector{Int64}, c5::Vector{Int64}, c6::Vector{Int64}})
   @ Base namedtuple.jl:263
 [3] getindex(t::Table{@NamedTuple{c1::String, c2::Int64, c3::Int64, c4::Int64, c5::Int64, c6::Int64}, 1, @NamedTuple{c1::Vector{String}, c2::Vector{Int64}, c3::Vector{Int64}, c4::Vector{Int64}, c5::Vector{Int64}, c6::Vector{Int64}}}, inds::Vector{Int64})
   @ TypedTables C:\Users\Thoma\.julia\packages\TypedTables\ItVth\src\Table.jl:318 [inlined]
 [4] filter(f::var"#main##0#main##1"{String}, a::Table{@NamedTuple{c1::String, c2::Int64, c3::Int64, c4::Int64, c5::Int64, c6::Int64}, 1, @NamedTuple{c1::Vector{String}, c2::Vector{Int64}, c3::Vector{Int64}, c4::Vector{Int64}, c5::Vector{Int64}, c6::Vector{Int64}}})
   @ Base array.jl:3011
 [5] main(args::Vector{String})
   @ Main C:\jul\JulWork\parajul\jct\test_jl5.jl:1556
 [6] _main(argc::Int32, argv::Ptr{Ptr{Int8}})
   @ Main C:\Julia-1.13.0-beta1\share\julia\juliac\juliac-buildscript.jl:35

  Verifier error #2: unresolved call from statement Base.map(f::Function, Core.tuple(Core.getfield(nt::@NamedTuple{c1::Vector{String}, c2::Vector{Int64}, c3::Vector{Int64}, c4::Vector{Int64}, c5::Vector{Int64}, c6::Vector{Int64}}, 2)::Vector{Int64}, Core.getfield(nt::@NamedTuple{c1::Vector{String}, c2::Vector{Int64}, c3::Vector{Int64}, c4::Vector{Int64}, c5::Vector{Int64}, c6::Vector{Int64}}, 3)::Vector{Int64}, Core.getfield(nt::@NamedTuple{c1::Vector{String}, c2::Vector{Int64}, c3::Vector{Int64}, c4::Vector{Int64}, c5::Vector{Int64}, c6::Vector{Int64}}, 4)::Vector{Int64}, Core.getfield(nt::@NamedTuple{c1::Vector{String}, c2::Vector{Int64}, c3::Vector{Int64}, c4::Vector{Int64}, c5::Vector{Int64}, c6::Vector{Int64}}, 5)::Vector{Int64}, Core.getfield(nt::@NamedTuple{c1::Vector{String}, c2::Vector{Int64}, c3::Vector{Int64}, c4::Vector{Int64}, c5::Vector{Int64}, c6::Vector{Int64}}, 6)::Vector{Int64})::NTuple{5, Vector{Int64}})::NTuple{5, Any}
  Stacktrace:
 [1] map(f::Function, t::Tuple{Vector{String}, Vararg{Vector{Int64}, 5}})
   @ Base tuple.jl:357 [inlined]
 [2] map(::Function, ::@NamedTuple{c1::Vector{String}, c2::Vector{Int64}, c3::Vector{Int64}, c4::Vector{Int64}, c5::Vector{Int64}, c6::Vector{Int64}})
   @ Base namedtuple.jl:263
 [3] getindex(t::Table{@NamedTuple{c1::String, c2::Int64, c3::Int64, c4::Int64, c5::Int64, c6::Int64}, 1, @NamedTuple{c1::Vector{String}, c2::Vector{Int64}, c3::Vector{Int64}, c4::Vector{Int64}, c5::Vector{Int64}, c6::Vector{Int64}}}, inds::Vector{Int64})
   @ TypedTables C:\Users\Thoma\.julia\packages\TypedTables\ItVth\src\Table.jl:318 [inlined]
 [4] filter(f::var"#main##0#main##1"{String}, a::Table{@NamedTuple{c1::String, c2::Int64, c3::Int64, c4::Int64, c5::Int64, c6::Int64}, 1, @NamedTuple{c1::Vector{String}, c2::Vector{Int64}, c3::Vector{Int64}, c4::Vector{Int64}, c5::Vector{Int64}, c6::Vector{Int64}}})
   @ Base array.jl:3011
 [5] main(args::Vector{String})
   @ Main C:\jul\JulWork\parajul\jct\test_jl5.jl:1556
 [6] _main(argc::Int32, argv::Ptr{Ptr{Int8}})
   @ Main C:\Julia-1.13.0-beta1\share\julia\juliac\juliac-buildscript.jl:35

2. Setting cells in the tables: causes compile errors, if the number of different column types is greater than 2:

using TypedTables

function (@main)(args::Vector{String})::Cint
  t = Table(c1 = ["1"], c2 = [1.2], c3 = [1])
  d = getfield(t, :data)
  r = 1
  for s in 1:3
    d[s][r] = getfield(d, s)[r]
  end
  return 0
end

It causes the compile errors

Verifier error #1: unresolved call from statement Base.convert((Core._svec_ref)((Core._compute_sparams)(setindex!(A::Array{T}, x, i::Int64) where T @ Base array.jl:1017::Method, Base.setindex!, Base.getfield(%new()::@NamedTuple{c1::Vector{String}, c2::Vector{Float64}, c3::Vector{Int64}}, φ ()::Int64)::Union{Vector{Float64}, Vector{Int64}, Vector{String}}, φ ()::Union{Float64, Int64, String}, 1)::Core.SimpleVector, 1)::Any, φ ()::Union{Float64, Int64, String})::Any
Stacktrace:
 [1] setindex!(A::Union{Vector{Float64}, Vector{Int64}, Vector{String}}, x::Union{Float64, Int64, String}, i::Int64)
   @ Base array.jl:1019 [inlined]
 [2] main(args::Vector{String})
   @ Main C:\jul\JulWork\parajul\jct\test_jl5.jl:1573
 [3] _main(argc::Int32, argv::Ptr{Ptr{Int8}})
   @ Main C:\Julia-1.13.0-beta1\share\julia\juliac\juliac-buildscript.jl:35

Verifier error #2: unresolved call from statement Base.convert((Core._svec_ref)((Core._compute_sparams)(setindex!(A::Array{T}, x, i::Int64) where T @ Base array.jl:1017::Method, Base.setindex!, Base.getfield(%new()::@NamedTuple{c1::Vector{String}, c2::Vector{Float64}, c3::Vector{Int64}}, φ ()::Int64)::Union{Vector{Float64}, Vector{Int64}, Vector{String}}, φ ()::Union{Float64, Int64, String}, 1)::Core.SimpleVector, 1)::Any, φ ()::Union{Float64, Int64, String})::Any
Stacktrace:
 [1] setindex!(A::Union{Vector{Float64}, Vector{Int64}, Vector{String}}, x::Union{Float64, Int64, String}, i::Int64)
   @ Base array.jl:1019 [inlined]
 [2] main(args::Vector{String})
   @ Main C:\jul\JulWork\parajul\jct\test_jl5.jl:1573
 [3] _main(argc::Int32, argv::Ptr{Ptr{Int8}})
   @ Main C:\Julia-1.13.0-beta1\share\julia\juliac\juliac-buildscript.jl:35