Polars dies on working code when POLARS_PANIC_ON_ERR=1 is set

[douglas-raillard-arm]

Checks

• I have checked that this issue has not already been reported.
• I have confirmed this bug exists on the latest version of Polars.

Reproducible example

import pandas as pd
print(pd.__version__)

import polars as pl
pl.show_versions()

df = pd.DataFrame(dict(a=["a"]), index=[1], dtype='category')
print(df)
pl_df = pl.from_pandas(df, include_index=True)
print(pl_df)

Log output

2.2.3
--------Version info---------
Polars:              1.17.1
Index type:          UInt32
Platform:            Linux-6.8.0-49-generic-x86_64-with-glibc2.39
Python:              3.12.3 (main, Nov  6 2024, 18:32:19) [GCC 13.2.0]
LTS CPU:             False

----Optional dependencies----
adbc_driver_manager  <not installed>
altair               <not installed>
boto3                1.35.48
cloudpickle          <not installed>
connectorx           <not installed>
deltalake            <not installed>
fastexcel            <not installed>
fsspec               <not installed>
gevent               <not installed>
google.auth          <not installed>
great_tables         <not installed>
matplotlib           <not installed>
nest_asyncio         1.6.0
numpy                2.1.2
openpyxl             <not installed>
pandas               2.2.3
pyarrow              17.0.0
pydantic             <not installed>
pyiceberg            <not installed>
sqlalchemy           <not installed>
torch                <not installed>
xlsx2csv             <not installed>
xlsxwriter           <not installed>
   a
1  a
thread '<unnamed>' panicked at /home/runner/work/polars/polars/crates/polars-error/src/lib.rs:45:37:
unexpected value while building Series of type Int64; found value of type String: "None"
stack backtrace:
   0: rust_begin_unwind
   1: core::panicking::panic_fmt
   2: <polars_error::ErrString as core::convert::From<T>>::from::panic_cold_display
   3: <polars_error::ErrString as core::convert::From<T>>::from
   4: polars_core::series::any_value::invalid_value_error
   5: polars_core::series::any_value::<impl polars_core::series::Series>::from_any_values_and_dtype
   6: polars_python::series::construction::<impl polars_python::series::PySeries>::new_from_any_values_and_dtype
   7: polars_python::series::construction::<impl polars_python::series::PySeries>::__pymethod_new_from_any_values_and_dtype__
   8: pyo3::impl_::trampoline::trampoline
   9: polars_python::series::construction::_::__INVENTORY::trampoline
  10: <unknown>
  11: _PyObject_MakeTpCall
  12: _PyEval_EvalFrameDefault
  13: _PyObject_Call_Prepend
  14: <unknown>
  15: <unknown>
  16: _PyObject_MakeTpCall
  17: _PyEval_EvalFrameDefault
  18: <unknown>
  19: _PyEval_EvalFrameDefault
  20: PyEval_EvalCode
  21: <unknown>
  22: <unknown>
  23: _PyRun_SimpleFileObject
  24: _PyRun_AnyFileObject
  25: Py_RunMain
  26: Py_BytesMain
  27: __libc_start_call_main
             at ./csu/../sysdeps/nptl/libc_start_call_main.h:58:16
  28: __libc_start_main_impl
             at ./csu/../csu/libc-start.c:360:3
  29: _start
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.
Traceback (most recent call last):
  File "/home/dourai01/Work/projects/lisa/.lisa-venv-3.12/lib/python3.12/site-packages/polars/_utils/construction/series.py", line 319, in _construct_series_with_fallbacks
    return constructor(name, values, strict)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
TypeError: 'str' object cannot be interpreted as an integer

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/home/dourai01/Work/projects/lisa/testcat.py", line 10, in <module>
    pl_df = pl.from_pandas(df, include_index=True)
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/dourai01/Work/projects/lisa/.lisa-venv-3.12/lib/python3.12/site-packages/polars/convert/general.py", line 587, in from_pandas
    pandas_to_pydf(
  File "/home/dourai01/Work/projects/lisa/.lisa-venv-3.12/lib/python3.12/site-packages/polars/_utils/construction/dataframe.py", line 1125, in pandas_to_pydf
    return arrow_to_pydf(
           ^^^^^^^^^^^^^^
  File "/home/dourai01/Work/projects/lisa/.lisa-venv-3.12/lib/python3.12/site-packages/polars/_utils/construction/dataframe.py", line 1238, in arrow_to_pydf
    df = df[names]
         ~~^^^^^^^
  File "/home/dourai01/Work/projects/lisa/.lisa-venv-3.12/lib/python3.12/site-packages/polars/dataframe/frame.py", line 1376, in __getitem__
    return get_df_item_by_key(self, key)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/dourai01/Work/projects/lisa/.lisa-venv-3.12/lib/python3.12/site-packages/polars/_utils/getitem.py", line 167, in get_df_item_by_key
    return _select_rows(df, key)  # type: ignore[arg-type]
           ^^^^^^^^^^^^^^^^^^^^^
  File "/home/dourai01/Work/projects/lisa/.lisa-venv-3.12/lib/python3.12/site-packages/polars/_utils/getitem.py", line 308, in _select_rows
    s = pl.Series("", key, dtype=Int64)
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/dourai01/Work/projects/lisa/.lisa-venv-3.12/lib/python3.12/site-packages/polars/series/series.py", line 289, in __init__
    self._s = sequence_to_pyseries(
              ^^^^^^^^^^^^^^^^^^^^^
  File "/home/dourai01/Work/projects/lisa/.lisa-venv-3.12/lib/python3.12/site-packages/polars/_utils/construction/series.py", line 144, in sequence_to_pyseries
    pyseries = _construct_series_with_fallbacks(
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/dourai01/Work/projects/lisa/.lisa-venv-3.12/lib/python3.12/site-packages/polars/_utils/construction/series.py", line 334, in _construct_series_with_fallbacks
    return PySeries.new_from_any_values_and_dtype(
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
pyo3_runtime.PanicException: unexpected value while building Series of type Int64; found value of type String: "None"

Issue description

When pl.from_pandas() is called on a pandas DataFrame with categorical column and POLARS_PANIC_ON_ERR=1, polars panics and raises a pyo3_runtime.PanicException. When the env var is set to 0, everything works as expected.

Expected behavior

Code that works under normal condition should keep working when POLARS_PANIC_ON_ERR=1, otherwise that reduces a lot its usability as a debugging tool as it will divert the attention away from the real problem being investigated.

On top of that, it also prevents setting it as an env var in e.g. a CI system, which could otherwise gather useful info for intermittent problems.

Installed versions

``` --------Version info--------- Polars: 1.17.1 Index type: UInt32 Platform: Linux-6.8.0-49-generic-x86_64-with-glibc2.39 Python: 3.12.3 (main, Nov 6 2024, 18:32:19) [GCC 13.2.0] LTS CPU: False

----Optional dependencies----
adbc_driver_manager
altair
boto3 1.35.48
cloudpickle
connectorx
deltalake
fastexcel
fsspec
gevent
google.auth
great_tables
matplotlib
nest_asyncio 1.6.0
numpy 2.1.2
openpyxl
pandas 2.2.3
pyarrow 17.0.0
pydantic
pyiceberg
sqlalchemy
torch
xlsx2csv
xlsxwriter

</details>

[douglas-raillard-arm]

The issue reported here also seems to come from that cause: #18896 (comment)

thread 'polars-0' panicked at /home/runner/work/polars/polars/crates/polars-error/src/lib.rs:45:37:
unsupported output type for dictionary packing: Duration(Nanosecond)

[cmdlineluser]

The example now runs for me on main. (I think thanks to #20248)

It seems previously there was a try / except involved in the parsing with a fallback in the except branch.

Setting the variable then prevented the fallback from running.