infinite loop when converting btw Pandas.DataFrame and DataFrames.DataFrame in v1.6.1

[yadongli]

The following creates an infinite loop and stack overflow in v1.6.1, it used to works fine in v1.6.0. The conversion btw Pandas and DataFrames are quite common use case, can you take a look?

using Pandas
using DataFrames

df = DataFrames.DataFrame(a=rand(10), b = rand(10))
df |> Pandas.DataFrame |> DataFrames.DataFrame

results:

StackOverflowError:

Stacktrace: [1] columns(x::Pandas.DataFrame) @ Tables ~/.julia/packages/Tables/M26tI/src/fallbacks.jl:252 [2] rows(x::Pandas.DataFrame) @ Tables ~/.julia/packages/Tables/M26tI/src/fallbacks.jl:88 --- the last 2 lines are repeated 39990 more times --- [79983] columns(x::Pandas.DataFrame) @ Tables ~/.julia/packages/Tables/M26tI/src/fallbacks.jl:252

am using [email protected] and [email protected]

[bkamins]

It seems to be unrelated to DataFrames.jl.

[yadongli]

Yes, you are right, Tables.rows(Pandas.DataFrame(df)) also overflows

[quinnj]

Fixed via https://github.com/JuliaPy/Pandas.jl/commit/b120a90654d942e2495eb711ed688e66ad1d6789?