[fix] No overload variant of "isna" matches argument type "NAType" [call-overload]

[LeMeteore]

• Update isna & notna overloads inside pandas-stubs/core/dtypes/missing.pyi

• Update test_isna function inside tests/test_pandas.py file

• Refs error: No overload variant of "isna" matches argument type "NAType" [call-overload] #264

• Tests updated

[LeMeteore]

When I tried to test with:

check(assert_type(pd.isna(pd.NA), bool), bool)                                                                                              
check(assert_type(pd.notna(pd.NA), bool), bool)

I had the following errors:

tests/test_pandas.py:130: error: Expression is of type "Any", not "bool"  [assert-type]
tests/test_pandas.py:130: error: No overload variant of "isna" matches argument type "NAType"  [call-overload]
tests/test_pandas.py:130: note: Possible overload variants:
tests/test_pandas.py:130: note:     def isna(obj: DataFrame) -> DataFrame
tests/test_pandas.py:130: note:     def isna(obj: Series[Any]) -> Series[bool]
tests/test_pandas.py:130: note:     def isna(obj: Union[Index, List[Any], Union[ExtensionArray, ndarray[Any, Any]]]) -> ndarray[Any, Any]
tests/test_pandas.py:130: note:     def isna(obj: Union[Union[str, bytes, date, datetime, timedelta, bool, complex, Timestamp, Timedelta], NaTType]) -> bool
tests/test_pandas.py:131: error: Expression is of type "Any", not "bool"  [assert-type]
tests/test_pandas.py:131: error: No overload variant of "notna" matches argument type "NAType"  [call-overload]
tests/test_pandas.py:131: note: Possible overload variants:
tests/test_pandas.py:131: note:     def notna(obj: DataFrame) -> DataFrame
tests/test_pandas.py:131: note:     def notna(obj: Series[Any]) -> Series[bool]
tests/test_pandas.py:131: note:     def notna(obj: Union[Index, List[Any], Union[ExtensionArray, ndarray[Any, Any]]]) -> ndarray[Any, Any]
tests/test_pandas.py:131: note:     def notna(obj: Union[Union[str, bytes, date, datetime, timedelta, bool, complex, Timestamp, Timedelta], NaTType]) -> bool

I decided to use np.nan instead of pd.NA:

check(assert_type(pd.isna(np.nan), bool), bool)                                                                                             
check(assert_type(pd.notna(np.nan), bool), bool) 

[Dr-Irv]

I decided to use np.nan instead of pd.NA:

I think you pulled the wrong type. You need to use NAType and NaTType. The first corresponds to pd.NA, the second to pd.NaT (which you should also test as well).

>>> import pandas as pd
>>> type(pd.NA)
<class 'pandas._libs.missing.NAType'>
>>> type(pd.NaT)
<class 'pandas._libs.tslibs.nattype.NaTType'>

[Dr-Irv]

This all looks good except for one small thing.

[Dr-Irv]

hit the wrong button - make the changes with the tests and we should be good

[Dr-Irv]

Thanks @LeMeteore

[LeMeteore]

Thanks @LeMeteore

Thanks to you all o/ I've learned something today.

[marcglobality]

re this one, I'm getting errors in mypy due pd.isna(None). Is this expected behaviour? Could we cover also None ?

[Dr-Irv]

re this one, I'm getting errors in mypy due pd.isna(None). Is this expected behaviour? Could we cover also None ?

Can you create an issue, with the example, and the output of mypy?

[marcglobality]

here you go #318