pandas-dev/pandas · error · TypeError

datetime64 type does not support operation '{how}'

Error message

datetime64 type does not support operation '{how}'

What it means

Raised by _groupby_op when the grouped array is datetime64 and the requested how is one of sum/prod/cumsum/cumprod/var/skew/kurt. Summing or multiplying absolute timestamps is not meaningful, so pandas rejects these reductions on groupby dispatch.

Source

Thrown at pandas/core/arrays/datetimelike.py:1624

    # ------------------------------------------------------------------
    # GroupBy Methods

    def _groupby_op(
        self,
        *,
        how: str,
        has_dropped_na: bool,
        min_count: int,
        ngroups: int,
        ids: npt.NDArray[np.intp],
        **kwargs,
    ):
        dtype = self.dtype
        if dtype.kind == "M":
            # Adding/multiplying datetimes is not valid
            if how in ["sum", "prod", "cumsum", "cumprod", "var", "skew", "kurt"]:
                raise TypeError(f"datetime64 type does not support operation '{how}'")
            if how in ["any", "all"]:
                # GH#34479
                raise TypeError(
                    f"'{how}' with datetime64 dtypes is no longer supported. "
                    f"Use (obj != pd.Timestamp(0)).{how}() instead."
                )

        elif isinstance(dtype, PeriodDtype):
            # Adding/multiplying Periods is not valid
            if how in ["sum", "prod", "cumsum", "cumprod", "var", "skew", "kurt"]:
                raise TypeError(f"Period type does not support {how} operations")
            if how in ["any", "all"]:
                # GH#34479
                raise TypeError(
                    f"'{how}' with PeriodDtype is no longer supported. "
                    f"Use (obj != pd.Period(0, freq)).{how}() instead."
                )
        # timedeltas we can add but not multiply

View on GitHub (pinned to 71959b8cb9)

Solutions

  1. Aggregate a numeric column instead, or convert the datetime column: use .min()/.max()/.median()/.count() which are supported.
  2. Compute a duration by subtracting a baseline first: (df['ts'] - df['ts'].min()).groupby(key).sum().
  3. Drop datetime columns before calling generic .sum()/.prod() reductions.
  4. Use .agg() with a function allow-list that excludes sum/prod for datetime dtypes.

Example fix

// before
df.groupby('key')['timestamp'].sum()  # TypeError: datetime64 type does not support operation 'sum'
// after
df.groupby('key')['timestamp'].agg(['min','max','count'])
Defensive patterns

Strategy: type-guard

Validate before calling

from pandas.api.types import is_datetime64_any_dtype
BAD = {'sum','prod','cumsum','cumprod','var','skew','kurt'}
if is_datetime64_any_dtype(col.dtype) and how in BAD:
    how = 'min'  # or skip the column

Type guard

def supports_groupby_op(col, how: str) -> bool:
    from pandas.api.types import is_datetime64_any_dtype
    BAD = {'sum','prod','cumsum','cumprod','var','skew','kurt'}
    return not (is_datetime64_any_dtype(col.dtype) and how in BAD)

Try / catch

try:
    out = df.groupby('key')['ts'].agg(how)
except TypeError as e:
    if 'datetime64 type does not support' in str(e):
        out = df.groupby('key')['ts'].agg(['min','max','count'])
    else:
        raise

Prevention

When it happens

Trigger: df.groupby(key)[ts_col].sum(), .prod(), .cumsum(), .var(), .skew(), .kurt() on a datetime64 column; reached via the branch at line 1621-1624.

Common situations: df.describe() or df.groupby().agg(['sum','mean']) applied blindly to datetime columns; pipelines that aggregate every numeric-looking column including timestamps.

Related errors

AI-assisted analysis of pandas-dev/pandas@71959b8cb9 (2026-08-07). Data as JSON: /api/errors/8998255087089b9b. Report an issue: GitHub.