What causes "Cannot cast NaN value to Integer dtype." in pandas-dev/pandas?

Use a Floating masked dtype (Float64) to preserve NaN, then cast after handling missing values.

pandas-dev/pandas · error · ValueError

Cannot cast NaN value to Integer dtype.

Error message

Cannot cast NaN value to Integer dtype.

What it means

Raised by _coerce_to_data_and_mask when the values are floating-point, the target is an Integer masked dtype, and the float data contains NaN that cannot be represented (under the non-pandas-NA mode where np.nan is not treated as the NA sentinel). Rather than silently wrapping NaN into an invalid integer bit pattern, pandas raises.

Source

Thrown at pandas/core/arrays/numeric.py:204

    if values.ndim != 1:
        raise TypeError("values must be a 1D list-like")

    if mask is None:
        if values.dtype.kind in "iu":
            # fastpath
            mask = np.zeros(len(values), dtype=np.bool_)
        elif values.dtype.kind == "f":
            # np.isnan is faster than is_numeric_na() for floats
            # github issue: #60066
            if is_nan_na():
                mask = np.isnan(values)
            else:
                mask = np.zeros(len(values), dtype=np.bool_)
                if dtype_cls.__name__.strip("_").startswith(("I", "U")):
                    wrong = np.isnan(values)
                    if wrong.any():
                        raise ValueError("Cannot cast NaN value to Integer dtype.")
        elif is_nan_na():
            mask = libmissing.is_numeric_na(values)
        else:
            # is_numeric_na will raise on non-numeric NAs
            libmissing.is_numeric_na(values)
            mask = libmissing.is_pdna_or_none(values)
    else:
        assert len(mask) == len(values)

    if mask.ndim != 1:
        raise TypeError("mask must be a 1D list-like")

    # infer dtype if needed
    if dtype is None:
        dtype = default_dtype
    else:
        dtype = dtype.numpy_dtype

View on GitHub (pinned to 71959b8cb9)

Solutions

Use a Floating masked dtype (Float64) to preserve NaN, then cast after handling missing values.
Drop or fill NaN before constructing: pd.array(np.nan_to_num(vals, nan=0.0), dtype='Int64').
Convert floats to the integer array via Series: pd.Series(vals).astype('Int64') after fillna.

Example fix

// before
pd.array([1.0, np.nan], dtype="Int64")  # raises: Cannot cast NaN value to Integer dtype.

// after
pd.array([1.0, np.nan], dtype="Float64")  # or fillna first

Defensive patterns

Strategy: validation

Validate before calling

import numpy as np

def to_int_masked_safe(vals):
    arr = np.asarray(vals, dtype="float64") if np.asarray(vals).dtype.kind == "f" else np.asarray(vals)
    if arr.dtype.kind == "f" and np.isnan(arr).any():
        raise ValueError("float source has NaN; use Float64 or fill NaN first")
    return arr

Type guard

def is_int_castable_float(vals) -> bool:
    import numpy as np
    arr = np.asarray(vals)
    return not (arr.dtype.kind == "f" and bool(np.isnan(arr).any()))

Try / catch

try:
    arr = pd.array(vals, dtype="Int64")
except ValueError as e:
    if "Cannot cast NaN value to Integer dtype" in str(e):
        arr = pd.array(vals, dtype="Float64")
    else:
        raise

Prevention

Use Float64 masked dtype when NaN may be present.
Fill/drop NaN before constructing integer masked arrays.
Centralize numeric coercion in a tested helper.

When it happens

Trigger: Constructing Int8/Int16/Int32/Int64 from a float numpy array that contains np.nan while the NA convention in use is not np.nan (is_nan_na() is False), e.g. pd.array([1.0, np.nan], dtype='Int64') in an environment configured with pd.NA as the sentinel.

Common situations: Mixed float source with NaN being pushed into an Integer dtype; library configuration that changed the NA sentinel behavior; data ingestion leaving NaN in numeric columns.

Related errors

AI-assisted analysis of pandas-dev/pandas@71959b8cb9 (2026-08-07). Data as JSON: /api/errors/2def657be00eafdc. Report an issue: GitHub.