df.dtypes fails with Arrow

[tvst]

Summary

Using df.dtypes to show the datatypes for a dataframe fails when you're using the Arrow codepath.

Steps to reproduce

Code snippet:

import streamlit as st
import pandas as pd

df = pd.read_json("https://cdn.jsdelivr.net/npm/vega-datasets@2/data/penguins.json")

st.write(df.dtypes)

Expected behavior:

The app should show a table with datatypes for the given dataframe

Actual behavior:

The app shows this exception:

Exception text

ArrowInvalid: ("Could not convert dtype('O') with type numpy.dtype[object_]: did not recognize Python value type when inferring an Arrow data type", 'Conversion failed for column 0 with type object')
Traceback:
File "/home/tvst/.local/share/virtualenvs/streamlit-vega-lite-demo-Se5mEOd4/lib/python3.9/site-packages/streamlit/script_runner.py", line 350, in _run_script
    exec(code, module.__dict__)
File "/media/storage/Projects/streamlit/streamlit-vega-lite-demo/demo.py", line 15, in <module>
    df.dtypes
File "/home/tvst/.local/share/virtualenvs/streamlit-vega-lite-demo-Se5mEOd4/lib/python3.9/site-packages/streamlit/__init__.py", line 478, in _transparent_write
    write(*args)
File "/home/tvst/.local/share/virtualenvs/streamlit-vega-lite-demo-Se5mEOd4/lib/python3.9/site-packages/streamlit/elements/write.py", line 182, in write
    self.dg.dataframe(arg)
File "/home/tvst/.local/share/virtualenvs/streamlit-vega-lite-demo-Se5mEOd4/lib/python3.9/site-packages/streamlit/elements/dataframe_selector.py", line 85, in dataframe
    return self.dg._arrow_dataframe(data, width, height)
File "/home/tvst/.local/share/virtualenvs/streamlit-vega-lite-demo-Se5mEOd4/lib/python3.9/site-packages/streamlit/elements/arrow.py", line 82, in _arrow_dataframe
    marshall(proto, data, default_uuid)
File "/home/tvst/.local/share/virtualenvs/streamlit-vega-lite-demo-Se5mEOd4/lib/python3.9/site-packages/streamlit/elements/arrow.py", line 160, in marshall
    proto.data = type_util.data_frame_to_bytes(df)
File "/home/tvst/.local/share/virtualenvs/streamlit-vega-lite-demo-Se5mEOd4/lib/python3.9/site-packages/streamlit/type_util.py", line 371, in data_frame_to_bytes
    table = pa.Table.from_pandas(df)
File "pyarrow/table.pxi", line 1561, in pyarrow.lib.Table.from_pandas
File "/home/tvst/.local/share/virtualenvs/streamlit-vega-lite-demo-Se5mEOd4/lib64/python3.9/site-packages/pyarrow/pandas_compat.py", line 594, in dataframe_to_arrays
    arrays = [convert_column(c, f)
File "/home/tvst/.local/share/virtualenvs/streamlit-vega-lite-demo-Se5mEOd4/lib64/python3.9/site-packages/pyarrow/pandas_compat.py", line 594, in <listcomp>
    arrays = [convert_column(c, f)
File "/home/tvst/.local/share/virtualenvs/streamlit-vega-lite-demo-Se5mEOd4/lib64/python3.9/site-packages/pyarrow/pandas_compat.py", line 581, in convert_column
    raise e
File "/home/tvst/.local/share/virtualenvs/streamlit-vega-lite-demo-Se5mEOd4/lib64/python3.9/site-packages/pyarrow/pandas_compat.py", line 575, in convert_column
    result = pa.array(col, type=type_, from_pandas=True, safe=safe)
File "pyarrow/array.pxi", line 302, in pyarrow.lib.array
File "pyarrow/array.pxi", line 83, in pyarrow.lib._ndarray_to_array
File "pyarrow/error.pxi", line 99, in pyarrow.lib.check_status

Is this a regression?

yes (arrow vs legacy)

Debug info

• Streamlit version: 0.86.0

[joelbs]

Similar issue. Anyone know anything about this?