Cache + Repeated conversion gives NoneType

[GabrieleLabanca]

Describe the bug
When using type hints and conversion to pa.typing.DataFrame[schema] multiple times, the code fails in a non-deterministic way, depending on how many times one calls the function. The error is TypeError: 'NoneType' object is not iterable.

• [ X ] I have checked that this issue has not already been reported.
• [ X ] I have confirmed this bug exists on the latest version of pandera. (0.21.1)
• (optional) I have confirmed this bug exists on the main branch of pandera.

Note: Please read this guide detailing how to provide the necessary information for us to reproduce your bug.

Code Sample, a copy-pastable example

Required installs: pytest, pandas, pandera. You may need to set N_ITER higher to get it to fail, in my case it ranged around 8-16.

from functools import cache

import pandas as pd
from pandera.typing import DataFrame
from pandera import DataFrameModel

import pytest

N_ITER = 15

class SchemaRegistryDevice(DataFrameModel):
    id: str
    categ: int


data = {
    "id": ["1", "2", "3"],
    "categ": [10, 20, 30],
}


@cache
def get_df_sensors_registry() -> DataFrame[SchemaRegistryDevice]:

    df_registry = pd.DataFrame(data).pipe(DataFrame[SchemaRegistryDevice])

    df_casted = DataFrame[SchemaRegistryDevice](df_registry)
    assert df_casted is not None
    return df_casted


def get_reg_test() -> DataFrame[SchemaRegistryDevice]:
    df_reg = get_df_sensors_registry()

    return DataFrame[SchemaRegistryDevice](df_reg)


@pytest.mark.parametrize("n", range(N_ITER))
def test_bikes_registry(n):
    reg = get_reg_test()

Expected behavior

In this example, the validations/casts are of course too many, but this reproduces (as minimally as I could) piping multiple functions. I would expect, however, that validating/casting to a pa.typing.DataFrame multiple times be idempotent. Even more surprisingly, this fails only after a certain number of calls to the same function (but not the same object), so I suppose this has to do with some hidden memoization inside the DataFrameModel.

Some thoughts:

• I could not reproduce it without caching, but I am not sure that this is the root cause, or only making the failure easier to happen (I need cache for my case anyway);
• the iteration at which it failed varies on my very machine: in VSCode is 14, in another terminal emulator is 9.

The command I run: python3 -m pytest main.py.
The full error:

============================================================================================================ test session starts =============================================================================================================
platform linux -- Python 3.13.0, pytest-8.3.4, pluggy-1.5.0
rootdir: /home/gabriele/Downloads/panderabug
plugins: typeguard-4.4.1
collected 15 items                                                                                                                                                                                                                           

main.py .........FFFFFF                                                                                                                                                                                                                [100%]

================================================================================================================== FAILURES ==================================================================================================================
___________________________________________________________________________________________________________ test_bikes_registry[9] ___________________________________________________________________________________________________________

n = 9

    @pytest.mark.parametrize("n", range(N_ITER))
    def test_bikes_registry(n):
>       reg = get_reg_test()

main.py:40: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
main.py:35: in get_reg_test
    return DataFrame[SchemaRegistryDevice](df_reg)
.venv/lib64/python3.13/site-packages/pandera/typing/common.py:127: in __patched_generic_alias_call
    result = self.__origin__(*args, **kwargs)
.venv/lib64/python3.13/site-packages/pandas/core/frame.py:712: in __init__
    data = data.copy(deep=False)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[TypeError("'NoneType' object is not iterable") raised in repr()] BlockManager object at 0x7fa850ab9910>, deep = False

    def copy(self, deep: bool | None | Literal["all"] = True) -> Self:
        """
        Make deep or shallow copy of BlockManager
    
        Parameters
        ----------
        deep : bool, string or None, default True
            If False or None, return a shallow copy (do not copy data)
            If 'all', copy data and a deep copy of the index
    
        Returns
        -------
        BlockManager
        """
        if deep is None:
            if using_copy_on_write():
                # use shallow copy
                deep = False
            else:
                # preserve deep copy for BlockManager with copy=None
                deep = True
    
        # this preserves the notion of view copying of axes
        if deep:
            # hit in e.g. tests.io.json.test_pandas
    
            def copy_func(ax):
                return ax.copy(deep=True) if deep == "all" else ax.view()
    
            new_axes = [copy_func(ax) for ax in self.axes]
        else:
            if using_copy_on_write():
                new_axes = [ax.view() for ax in self.axes]
            else:
>               new_axes = list(self.axes)
E               TypeError: 'NoneType' object is not iterable

.venv/lib64/python3.13/site-packages/pandas/core/internals/managers.py:591: TypeError
__________________________________________________________________________________________________________ test_bikes_registry[10] ___________________________________________________________________________________________________________

n = 10

    @pytest.mark.parametrize("n", range(N_ITER))
    def test_bikes_registry(n):
>       reg = get_reg_test()

main.py:40: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
main.py:35: in get_reg_test
    return DataFrame[SchemaRegistryDevice](df_reg)
.venv/lib64/python3.13/site-packages/pandera/typing/common.py:127: in __patched_generic_alias_call
    result = self.__origin__(*args, **kwargs)
.venv/lib64/python3.13/site-packages/pandas/core/frame.py:712: in __init__
    data = data.copy(deep=False)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[TypeError("'NoneType' object is not iterable") raised in repr()] BlockManager object at 0x7fa850ab9910>, deep = False

    def copy(self, deep: bool | None | Literal["all"] = True) -> Self:
        """
        Make deep or shallow copy of BlockManager
    
        Parameters
        ----------
        deep : bool, string or None, default True
            If False or None, return a shallow copy (do not copy data)
            If 'all', copy data and a deep copy of the index
    
        Returns
        -------
        BlockManager
        """
        if deep is None:
            if using_copy_on_write():
                # use shallow copy
                deep = False
            else:
                # preserve deep copy for BlockManager with copy=None
                deep = True
    
        # this preserves the notion of view copying of axes
        if deep:
            # hit in e.g. tests.io.json.test_pandas
    
            def copy_func(ax):
                return ax.copy(deep=True) if deep == "all" else ax.view()
    
            new_axes = [copy_func(ax) for ax in self.axes]
        else:
            if using_copy_on_write():
                new_axes = [ax.view() for ax in self.axes]
            else:
>               new_axes = list(self.axes)
E               TypeError: 'NoneType' object is not iterable

.venv/lib64/python3.13/site-packages/pandas/core/internals/managers.py:591: TypeError
__________________________________________________________________________________________________________ test_bikes_registry[11] ___________________________________________________________________________________________________________

n = 11

    @pytest.mark.parametrize("n", range(N_ITER))
    def test_bikes_registry(n):
>       reg = get_reg_test()

main.py:40: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
main.py:35: in get_reg_test
    return DataFrame[SchemaRegistryDevice](df_reg)
.venv/lib64/python3.13/site-packages/pandera/typing/common.py:127: in __patched_generic_alias_call
    result = self.__origin__(*args, **kwargs)
.venv/lib64/python3.13/site-packages/pandas/core/frame.py:712: in __init__
    data = data.copy(deep=False)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[TypeError("'NoneType' object is not iterable") raised in repr()] BlockManager object at 0x7fa850ab9910>, deep = False

    def copy(self, deep: bool | None | Literal["all"] = True) -> Self:
        """
        Make deep or shallow copy of BlockManager
    
        Parameters
        ----------
        deep : bool, string or None, default True
            If False or None, return a shallow copy (do not copy data)
            If 'all', copy data and a deep copy of the index
    
        Returns
        -------
        BlockManager
        """
        if deep is None:
            if using_copy_on_write():
                # use shallow copy
                deep = False
            else:
                # preserve deep copy for BlockManager with copy=None
                deep = True
    
        # this preserves the notion of view copying of axes
        if deep:
            # hit in e.g. tests.io.json.test_pandas
    
            def copy_func(ax):
                return ax.copy(deep=True) if deep == "all" else ax.view()
    
            new_axes = [copy_func(ax) for ax in self.axes]
        else:
            if using_copy_on_write():
                new_axes = [ax.view() for ax in self.axes]
            else:
>               new_axes = list(self.axes)
E               TypeError: 'NoneType' object is not iterable

.venv/lib64/python3.13/site-packages/pandas/core/internals/managers.py:591: TypeError
__________________________________________________________________________________________________________ test_bikes_registry[12] ___________________________________________________________________________________________________________

n = 12

    @pytest.mark.parametrize("n", range(N_ITER))
    def test_bikes_registry(n):
>       reg = get_reg_test()

main.py:40: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
main.py:35: in get_reg_test
    return DataFrame[SchemaRegistryDevice](df_reg)
.venv/lib64/python3.13/site-packages/pandera/typing/common.py:127: in __patched_generic_alias_call
    result = self.__origin__(*args, **kwargs)
.venv/lib64/python3.13/site-packages/pandas/core/frame.py:712: in __init__
    data = data.copy(deep=False)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[TypeError("'NoneType' object is not iterable") raised in repr()] BlockManager object at 0x7fa850ab9910>, deep = False

    def copy(self, deep: bool | None | Literal["all"] = True) -> Self:
        """
        Make deep or shallow copy of BlockManager
    
        Parameters
        ----------
        deep : bool, string or None, default True
            If False or None, return a shallow copy (do not copy data)
            If 'all', copy data and a deep copy of the index
    
        Returns
        -------
        BlockManager
        """
        if deep is None:
            if using_copy_on_write():
                # use shallow copy
                deep = False
            else:
                # preserve deep copy for BlockManager with copy=None
                deep = True
    
        # this preserves the notion of view copying of axes
        if deep:
            # hit in e.g. tests.io.json.test_pandas
    
            def copy_func(ax):
                return ax.copy(deep=True) if deep == "all" else ax.view()
    
            new_axes = [copy_func(ax) for ax in self.axes]
        else:
            if using_copy_on_write():
                new_axes = [ax.view() for ax in self.axes]
            else:
>               new_axes = list(self.axes)
E               TypeError: 'NoneType' object is not iterable

.venv/lib64/python3.13/site-packages/pandas/core/internals/managers.py:591: TypeError
__________________________________________________________________________________________________________ test_bikes_registry[13] ___________________________________________________________________________________________________________

n = 13

    @pytest.mark.parametrize("n", range(N_ITER))
    def test_bikes_registry(n):
>       reg = get_reg_test()

main.py:40: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
main.py:35: in get_reg_test
    return DataFrame[SchemaRegistryDevice](df_reg)
.venv/lib64/python3.13/site-packages/pandera/typing/common.py:127: in __patched_generic_alias_call
    result = self.__origin__(*args, **kwargs)
.venv/lib64/python3.13/site-packages/pandas/core/frame.py:712: in __init__
    data = data.copy(deep=False)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[TypeError("'NoneType' object is not iterable") raised in repr()] BlockManager object at 0x7fa850ab9910>, deep = False

    def copy(self, deep: bool | None | Literal["all"] = True) -> Self:
        """
        Make deep or shallow copy of BlockManager
    
        Parameters
        ----------
        deep : bool, string or None, default True
            If False or None, return a shallow copy (do not copy data)
            If 'all', copy data and a deep copy of the index
    
        Returns
        -------
        BlockManager
        """
        if deep is None:
            if using_copy_on_write():
                # use shallow copy
                deep = False
            else:
                # preserve deep copy for BlockManager with copy=None
                deep = True
    
        # this preserves the notion of view copying of axes
        if deep:
            # hit in e.g. tests.io.json.test_pandas
    
            def copy_func(ax):
                return ax.copy(deep=True) if deep == "all" else ax.view()
    
            new_axes = [copy_func(ax) for ax in self.axes]
        else:
            if using_copy_on_write():
                new_axes = [ax.view() for ax in self.axes]
            else:
>               new_axes = list(self.axes)
E               TypeError: 'NoneType' object is not iterable

.venv/lib64/python3.13/site-packages/pandas/core/internals/managers.py:591: TypeError
__________________________________________________________________________________________________________ test_bikes_registry[14] ___________________________________________________________________________________________________________

n = 14

    @pytest.mark.parametrize("n", range(N_ITER))
    def test_bikes_registry(n):
>       reg = get_reg_test()

main.py:40: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
main.py:35: in get_reg_test
    return DataFrame[SchemaRegistryDevice](df_reg)
.venv/lib64/python3.13/site-packages/pandera/typing/common.py:127: in __patched_generic_alias_call
    result = self.__origin__(*args, **kwargs)
.venv/lib64/python3.13/site-packages/pandas/core/frame.py:712: in __init__
    data = data.copy(deep=False)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[TypeError("'NoneType' object is not iterable") raised in repr()] BlockManager object at 0x7fa850ab9910>, deep = False

    def copy(self, deep: bool | None | Literal["all"] = True) -> Self:
        """
        Make deep or shallow copy of BlockManager
    
        Parameters
        ----------
        deep : bool, string or None, default True
            If False or None, return a shallow copy (do not copy data)
            If 'all', copy data and a deep copy of the index
    
        Returns
        -------
        BlockManager
        """
        if deep is None:
            if using_copy_on_write():
                # use shallow copy
                deep = False
            else:
                # preserve deep copy for BlockManager with copy=None
                deep = True
    
        # this preserves the notion of view copying of axes
        if deep:
            # hit in e.g. tests.io.json.test_pandas
    
            def copy_func(ax):
                return ax.copy(deep=True) if deep == "all" else ax.view()
    
            new_axes = [copy_func(ax) for ax in self.axes]
        else:
            if using_copy_on_write():
                new_axes = [ax.view() for ax in self.axes]
            else:
>               new_axes = list(self.axes)
E               TypeError: 'NoneType' object is not iterable

.venv/lib64/python3.13/site-packages/pandas/core/internals/managers.py:591: TypeError
============================================================================================================== warnings summary ==============================================================================================================
main.py::test_bikes_registry[0]
main.py::test_bikes_registry[0]
main.py::test_bikes_registry[0]
main.py::test_bikes_registry[0]
main.py::test_bikes_registry[0]
main.py::test_bikes_registry[0]
  /home/gabriele/Downloads/panderabug/.venv/lib64/python3.13/site-packages/multimethod/__init__.py:453: DeprecationWarning: use `parametric(<base>, <func>)` as a type instead
    warnings.warn("use `parametric(<base>, <func>)` as a type instead", DeprecationWarning)

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
========================================================================================================== short test summary info ===========================================================================================================
FAILED main.py::test_bikes_registry[9] - TypeError: 'NoneType' object is not iterable
FAILED main.py::test_bikes_registry[10] - TypeError: 'NoneType' object is not iterable
FAILED main.py::test_bikes_registry[11] - TypeError: 'NoneType' object is not iterable
FAILED main.py::test_bikes_registry[12] - TypeError: 'NoneType' object is not iterable
FAILED main.py::test_bikes_registry[13] - TypeError: 'NoneType' object is not iterable
FAILED main.py::test_bikes_registry[14] - TypeError: 'NoneType' object is not iterable
================================================================================================== 6 failed, 9 passed, 6 warnings in 1.20s ===================================================================================================

Desktop (please complete the following information):

• OS: Fedora Linux 41
• Version: Pandera (0.21.1)