search

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

Copy-paste old arrow code into subdir to remove Arrow package dependency #145

Closed quinnj closed 3 years ago

quinnj commented 3 years ago

Hmmm, we seem to have bigger compat problems; the arrow code is doing getlevels(x::CategoricalArray) = x.pool.index which seems to not be valid in newer CategoricalArrays releases; but if we put upper bounds on CAs.jl, then that makes the DataFrames.jl bump not work. @nalimilan any advice?

quinnj commented 3 years ago

Ok, looks like we should just be calling levels(c); so let's see how CI likes this

nalimilan commented 3 years ago

Failures on 32-bit Windows seem to be due to passing an Int64 to fieldtype. That's weird since Tables.columnindex seems to always return an Int, but maybe just do fieldtype(T, Int(col)).

nalimilan commented 3 years ago

32-bit Windows still fails, but looking at the AppVeyor log, the fieldtype failure has been there for at least one year so this PR doesn't make things worse.

nalimilan commented 3 years ago

Should we tag a release now?

quinnj commented 3 years ago

Tagged