JuliaData / Feather.jl

Read and write feather files in pure Julia
https://juliadata.github.io/Feather.jl/stable
Other
109 stars 27 forks source link

Is that possible to support DataFrame with tuple type? #130

Open reactgary opened 4 years ago

reactgary commented 4 years ago

Something like this? It will be quite useful in some cases...

a = DataFrame(f1 = Float32[], f2 = Float32[], f3 = Tuple{Float32, Float32}[])

ExpandingMan commented 4 years ago

I am currently pretty far into completely rewriting Arrow.jl, and yes, in principle the arrow format does support this. I'm not sure I'll have these on the very first release, but it's definitely something I'd like to have before too long. I'm pretty sure the feather metadata can't support this though (roughly speaking, feather is a subset of arrow), so even when I have that you'll have to use the arrow file format rather than feather.