pandas-dev/pandas · error · ValueError

NumpyExtensionArray must be 1-dimensional.

Error message

NumpyExtensionArray must be 1-dimensional.

What it means

Raised by NumpyExtensionArray.__init__ when values.ndim == 0 (a scalar 0-dimensional ndarray). NumpyExtensionArray is documented as 1-dimensional; a 0-D input has no length and cannot back an ExtensionArray. (2-D is technically accepted but not advertised, per the inline comment.) The check is on ndim==0 specifically.

Solutions

1. Wrap the scalar in a 1-element 1-D array: np.atleast_1d(np.asarray(value)).
2. Use np.array([value]) explicitly to get a 1-D length-1 array.
3. Reconsider whether an ExtensionArray is the right container for a single scalar.

Example fix

# before
pd.arrays.NumpyExtensionArray(np.array(5))  # raises (0-D)

# after
pd.arrays.NumpyExtensionArray(np.atleast_1d(np.array(5)))

import numpy as np

def to_1d_numpy_extension_array(values):
    arr = np.asarray(values)
    if arr.ndim == 0:
        arr = np.atleast_1d(arr)
    return pd.arrays.NumpyExtensionArray(arr)

import numpy as np

def is_nonzero_nd(obj) -> bool:
    return hasattr(obj, 'ndim') and obj.ndim >= 1

try:
    pd.arrays.NumpyExtensionArray(values)
except ValueError as e:
    if '1-dimensional' in str(e):
        pd.arrays.NumpyExtensionArray(np.atleast_1d(np.asarray(values)))
    else:
        raise

When it happens

Trigger: pd.arrays.NumpyExtensionArray(np.array(5)) — np.array(5) is 0-D. Passing the result of np.asarray(scalar) or np.float64(1.0). Extracting a scalar via .item()-like ops and feeding it back.

Common situations: Code that indexes a single element and keeps it as a 0-D array (e.g. arr[arr > 0].sum() shape) then tries to wrap it; arithmetic producing a scalar numpy value.

Related errors

• Cannot construct from scalar data. Pass a sequence instead.
• 'values' must be a NumPy array, not
• Array with ndim > 2 is not supported.
• Cannot create a from a MultiIndex.
• cannot evaluate scalar only bool ops

AI-assisted analysis of pandas-dev/pandas@3b7651241d (2026-08-11). Data as JSON: /api/errors/8e949984b3d9d01e. Report an issue: GitHub.

Appendix: source

Thrown at pandas/core/arrays/numpy_.py:134

    _dtype: NumpyEADtype
    _internal_fill_value = np.nan

    # ------------------------------------------------------------------------
    # Constructors

    def __init__(
        self, values: np.ndarray | NumpyExtensionArray, copy: bool = False
    ) -> None:
        if isinstance(values, type(self)):
            values = values._ndarray
        if not isinstance(values, np.ndarray):
            raise ValueError(
                f"'values' must be a NumPy array, not {type(values).__name__}"
            )

        if values.ndim == 0:
            # Technically we support 2, but do not advertise that fact.
            raise ValueError("NumpyExtensionArray must be 1-dimensional.")

        if copy:
            values = values.copy()

        dtype = NumpyEADtype(values.dtype)
        super().__init__(values, dtype)

    @classmethod
    def _from_sequence(
        cls, scalars, *, dtype: Dtype | None = None, copy: bool = False
    ) -> NumpyExtensionArray:
        if isinstance(dtype, NumpyEADtype):
            dtype = dtype._dtype
        if dtype is not None:
            dtype = np.dtype(dtype)  # type: ignore[arg-type]

        if dtype is not None and dtype.kind in "iu":
            # GH#41724 - validate NaN before casting float -> int

View on GitHub (pinned to 3b7651241d)