pandas-dev/pandas · error · ValueError

dtype ' ' is invalid, should be np.timedelta64 dtype

Error message

dtype '{dtype}' is invalid, should be np.timedelta64 dtype

What it means

Raised by _validate_td64_dtype when the resolved dtype is not a numpy timedelta64 dtype at all (after the unit-less 'm8' case is handled separately). This catches inputs like 'int64', 'float64', 'datetime64[ns]', or arbitrary object/category dtypes passed where a timedelta dtype is required.

Solutions

1. Pass a valid numpy timedelta64 dtype (e.g. 'timedelta64[ns]').
2. If you have integer data, convert it via `pd.to_timedelta(..., unit=...)` instead of forcing a timedelta dtype.
3. Validate that `pandas_dtype(dtype).kind == 'm'` before constructing.

Example fix

# before
pd.TimedeltaIndex([1, 2, 3], dtype='int64')  # ValueError

# after
pd.TimedeltaIndex([1, 2, 3], dtype='timedelta64[ns]', unit='s')

import numpy as np
from pandas.api.types import pandas_dtype

def is_valid_td_dtype(dtype) -> bool:
    dt = pandas_dtype(dtype)
    return dt.kind == 'm' and dt != np.dtype('m8')

null

try:
    idx = pd.TimedeltaIndex(data, dtype=dtype)
except ValueError as e:
    if 'should be np.timedelta64' in str(e):
        idx = pd.TimedeltaIndex(data, dtype='timedelta64[ns]')
    else:
        raise

When it happens

Trigger: `pd.TimedeltaIndex([], dtype='int64')`, or passing a datetime64/category/object dtype string to a timedelta-validated constructor.

Common situations: Reusing a dtype variable intended for a different column; dynamically building dtype strings and supplying the wrong kind.

Related errors

• Passing in 'timedelta' dtype with no precision is not…
• Supported timedelta64 resolutions are 's', 'ms', 'us', 'ns'
• Cannot divide vectors with unequal lengths
• Cannot multiply with
• Cannot multiply with unequal lengths

AI-assisted analysis of pandas-dev/pandas@3b7651241d (2026-08-11). Data as JSON: /api/errors/0ef22e6f32ed31a6. Report an issue: GitHub.

Appendix: source

Thrown at pandas/core/arrays/timedeltas.py:1436

    # coerce Index to np.ndarray, converting string-dtype if necessary
    values = np.asarray(data, dtype=np.object_)

    result = array_to_timedelta64(values, unit=unit, errors=errors)
    return result


def _validate_td64_dtype(dtype) -> DtypeObj:
    dtype = pandas_dtype(dtype)
    if dtype == np.dtype("m8"):
        # no precision disallowed GH#24806
        msg = (
            "Passing in 'timedelta' dtype with no precision is not allowed. "
            "Please pass in 'timedelta64[ns]' instead."
        )
        raise ValueError(msg)

    if not lib.is_np_dtype(dtype, "m"):
        raise ValueError(f"dtype '{dtype}' is invalid, should be np.timedelta64 dtype")
    elif not is_supported_dtype(dtype):
        raise ValueError("Supported timedelta64 resolutions are 's', 'ms', 'us', 'ns'")

    return dtype

View on GitHub (pinned to 3b7651241d)