sample doesn't preserve original object type

[MarcoGorelli]

What happened?

import polars as pl
import pandas as pd
import altair as alt

df_pd = pd.DataFrame({'a': [1,2,3]})
print(alt.sample(df_pd, n=2))
df_pl = pl.DataFrame({'a': [1,2,3]})
print(alt.sample(df_pl, n=2))

This outputs (exact values differ between runs due to the nature of sample, but the points is - Polars.DataFrame gets transformed to PyArrow)

   a
0  1
1  2
pyarrow.Table
a: int64
----
a: [[1,2]]

What would you like to happen instead?

The docs for alt.sample say

Reduce the size of the data model by sampling without replacement.

so I'd have expected the input library to be preserved, and to get something like:

   a
1  1
2  2
shape: (2, 1)
┌─────┐
│ a   │
│ --- │
│ i64 │
╞═════╡
│ 1   │
│ 2   │
└─────┘

Which version of Altair are you using?

5.3.0