"For" loop over enumeration leads to TypeError

[dominusmi]

Hello,

for (i, category) in enumerate( skipmissing(train_df[:category_name] ) )
    #code
end

Leads to TypeError

TypeError: next: in typeassert, expected WeakRefString{UInt8}, got String

Stacktrace:
 [1] next(::Missings.EachSkipMissing{WeakRefStrings.WeakRefStringArray{Union{Missings.Missing, WeakRefString{UInt8}},1}}, ::Int64) at /home/edoardo/.julia/v0.6/Missings/src/Missings.jl:363
 [2] next at ./iterators.jl:65 [inlined]
 [3] anonymous at ./<missing>:?
 [4] include_string(::String, ::String) at ./loading.jl:522

where train_df is a DataFrame (DataFrames.jl) and the type of :category_name column is of type

WeakRefStrings.WeakRefStringArray{Union{Missings.Missing, WeakRefString{UInt8}},1}

I don't know if I'm using it wrong or if there is an actual issue.
For the moment my workaround is to not use skipmissing directly:

for (i, category) in enumerate( train_df[:category_name] )
    ismissing(category)
end

but from the source code which I tried to read, it seems skipmissing somehow optimised this step, ergo why I'm curious

Have a good day!

[nalimilan]

I think that's due to the fact that WeakRefStringArray used to lie about its element type. It should be fixed by JuliaData/WeakRefStrings.jl#17. Closing since it's not a Missings.jl bug, but feel free to file it against WeakRefStrings if it persists with the next release.