pandas-dev/pandas · error · TypeError

' ' with dtype does not support operation ' ' with pyarrow…

Error message

'{type(self).__name__}' with dtype {self.dtype} does not support operation '{name}' with pyarrow version {pa.__version__}. '{name}' may be supported by upgrading pyarrow.

What it means

The companion guard to 146: when pyarrow.compute has the named function but invoking it raises AttributeError, NotImplementedError, or TypeError (typically because the installed pyarrow lacks a kernel for this dtype), pandas reports that the operation may be supported by upgrading pyarrow. The current and operation names are included to guide the upgrade.

Solutions

  1. Upgrade pyarrow to the latest version (pip install -U pyarrow).
  2. Cast to numpy dtype and compute the statistic with scipy/numpy instead.
  3. Pin pandas and pyarrow to compatible versions per pandas release notes.

Example fix

// before
# pyarrow too old
s = pd.Series([1., 2., 3.], dtype="double[pyarrow]")
s.kurt()
// after
pip install -U pyarrow
s = pd.Series([1., 2., 3.], dtype="double[pyarrow]")
s.kurt()
Defensive patterns

Strategy: retry

Validate before calling

import pyarrow as pa

def pyarrow_supports(name, min_version="10.0.0") -> bool:
    from packaging.version import Version
    return Version(pa.__version__) >= Version(min_version)

Type guard

import pyarrow as pa

def pyarrow_recent_enough(min_version="12.0.0") -> bool:
    from packaging.version import Version
    return Version(pa.__version__) >= Version(min_version)

Try / catch

try:
    s.kurt()
except TypeError as e:
    if "upgrading pyarrow" in str(e):
        s.astype("float64").kurt()  # fallback before upgrade
    else:
        raise

Prevention

When it happens

Trigger: Calling a reduction (e.g. .skew(), .kurt(), .sem(), .median()) whose pyarrow.compute kernel exists but fails for the array's pyarrow type, on an older pyarrow install.

Common situations: CI or production environments pinning an older pyarrow; user upgrades pandas without upgrading pyarrow; new reduction added to pandas not yet available in the user's pyarrow.

Related errors

AI-assisted analysis of pandas-dev/pandas@3b7651241d (2026-08-11). Data as JSON: /api/errors/36496672df97abe4. Report an issue: GitHub.

Appendix: source

Thrown at pandas/core/arrays/arrow/array.py:2625

        elif name == "median":
            # GH 52679: Use quantile instead of approximate_median
            kwargs["q"] = 0.5
        elif name in ["std", "var", "sem"] and "ddof" not in kwargs:
            # pyarrow defaults to ddof=0, pandas behavior is ddof=1
            kwargs["ddof"] = 1
        elif name in ["skew", "kurt"] and "biased" not in kwargs:
            kwargs["biased"] = False

        try:
            result = pyarrow_meth(data_to_reduce, skip_nulls=skipna, **kwargs)
        except (AttributeError, NotImplementedError, TypeError) as err:
            msg = (
                f"'{type(self).__name__}' with dtype {self.dtype} "
                f"does not support operation '{name}' with pyarrow "
                f"version {pa.__version__}. '{name}' may be supported by "
                f"upgrading pyarrow."
            )
            raise TypeError(msg) from err
        if name == "median":
            # GH 52679: Use quantile instead of approximate_median; returns array
            result = result[0]

        if name in ["min", "max", "sum"] and pa.types.is_duration(pa_type):
            result = result.cast(pa_type)
        if name in ["median", "mean"] and pa.types.is_temporal(pa_type):
            nbits = pa_type.bit_width
            if nbits == 32:
                result = result.cast(pa.int32(), safe=False)
            else:
                result = result.cast(pa.int64(), safe=False)
            result = result.cast(pa_type)
        if name in ["std", "sem"] and pa.types.is_temporal(pa_type):
            result = result.cast(pa.int64(), safe=False)
            if pa.types.is_duration(pa_type):
                result = result.cast(pa_type)
            elif pa.types.is_time(pa_type):

View on GitHub (pinned to 3b7651241d)