conflict with CSV and GNNGraphs when running Flux.batch

[doyku]

The issue comes from
sort(collect(keys(x1))) == sort(collect(keys(x2))) || @error "cannot concatenate feature data with different keys"

because Julia does not know which version of defalg to use for the sorting algorithm since InlineStrings (loaded by CSV) has a version of defalg that is compatible with this line.

Adding this line: Base.Sort.defalg(x::AbstractArray{<:Missing}) = QuickSort
resolved the ambiguity and allowed Flux.batch to run

[oysteinsolheim]

see #98

[doyku]

ok. Given that issue is closed should I expect to add that line in the future or will a future version fix it?

[CarloLucibello]

Technically this should be fixed in julia itself (see JuliaLang/julia#43426), but since that may take long and this problem pops up frequently we better add a workaround here. Unfortunately, we cannot define Base.Sort.defalg(x::AbstractArray{<:Missing}) = QuickSort in this repo, since it would be piracy. I'll try to figure something out

[doyku]

Not sure what you mean by it being Piracy. It looks like Inline strings has this sort of line:
https://github.com/JuliaStrings/InlineStrings.jl/blob/main/src/InlineStrings.jl#L698
Base.Sort.defalg(::AbstractArray{<:Union{SmallInlineStrings, Missing}}) = InlineStringSort

[CarloLucibello]

Not sure what you mean by it being Piracy. I

https://docs.julialang.org/en/v1/manual/style-guide/#:~:text=%22Type%20piracy%22%20refers%20to%20the,piracy%20with%20little%20ill%20effect.