pandas-dev/pandas · error · ValueError

Column is backed by an extension array, which is not…

Error message

Column {colname} is backed by an extension array, which is not supported by the numba engine.

What it means

Raised by `validate_values_for_numba` for a column whose dtype is an extension array (e.g. `Int64`, `Float64`, `boolean`, nullable types). Even if the dtype is numeric, numba operates on plain numpy buffers and cannot consume pandas ExtensionArrays, so pandas rejects them up front and names the offending column.

Solutions

Cast extension columns to numpy equivalents: `df[col] = df[col].astype('int64')` (handle NA first).
Use `df.select_dtypes(exclude='extension')` or filter out EA columns before the numba call.
Fall back to the python engine for frames that must keep nullable dtypes.

Example fix

// before
df.apply(func, engine='numba')  # df has Int64 (nullable) column
// after
df.astype({'col': 'int64'}).apply(func, engine='numba')

Defensive patterns

Strategy: validation

Validate before calling

def no_ea_numba_apply(df, func, **kw):
    ea_cols = [c for c, d in df.dtypes.items() if pd.api.types.is_extension_array_dtype(d)]
    if ea_cols:
        raise ValueError(f'Extension-array columns block numba: {ea_cols}')
    return df.apply(func, engine='numba', **kw)

Type guard

def frame_has_no_extension_arrays(df) -> bool:
    return not any(pd.api.types.is_extension_array_dtype(d) for d in df.dtypes)

Try / catch

try:
    out = df.apply(func, engine='numba')
except ValueError as e:
    if 'extension array' in str(e):
        cast = df.copy()
        for c in cast.columns:
            if pd.api.types.is_extension_array_dtype(cast[c]):
                cast[c] = cast[c].astype(cast[c].dtype._subtype if hasattr(cast[c].dtype, '_subtype') else 'float64')
        out = cast.apply(func, engine='numba')
    else:
        raise

Prevention

Avoid convert_dtypes() before numba paths.
Cast nullable Int64/Float64 to numpy int64/float64 after handling NaNs.
Document which pipelines require plain numpy dtypes.

When it happens

Trigger: `df.apply(func, engine='numba')` where any column is a nullable/extension dtype (`'Int64'`, `'Float64'`, `'boolean'`, etc.). The check `is_extension_array_dtype(dtype)` fires after the numeric check, so it only triggers for numeric extension dtypes.

Common situations: Frames produced by `convert_dtypes()` (which yields nullable Int64/Float64/boolean), or by reading data with `dtype='Int64'`; migrating to nullable dtypes without realizing numba does not support them.

Related errors

AI-assisted analysis of pandas-dev/pandas@3b7651241d (2026-08-11). Data as JSON: /api/errors/612fe6aa6af0d5c1. Report an issue: GitHub.

Appendix: source

Thrown at pandas/core/apply.py:984

    def generate_numba_apply_func(
        func, nogil: bool = True, parallel: bool = False
    ) -> Callable[[npt.NDArray, Index, Index], dict[int, Any]]:
        pass

    @abc.abstractmethod
    def apply_with_numba(self):
        pass

    def validate_values_for_numba(self) -> None:
        # Validate column dtypes all OK
        for colname, dtype in self.obj.dtypes.items():
            if not is_numeric_dtype(dtype):
                raise ValueError(
                    f"Column {colname} must have a numeric dtype. "
                    f"Found '{dtype}' instead"
                )
            if is_extension_array_dtype(dtype):
                raise ValueError(
                    f"Column {colname} is backed by an extension array, "
                    f"which is not supported by the numba engine."
                )

    @abc.abstractmethod
    def wrap_results_for_axis(
        self, results: ResType, res_index: Index
    ) -> DataFrame | Series:
        pass

    # ---------------------------------------------------------------

    @property
    def res_columns(self) -> Index:
        return self.result_columns

    @property
    def columns(self) -> Index:

View on GitHub (pinned to 3b7651241d)