pola-rs/polars · error · NotImplementedError

functionality for `nan_as_null` has not been implemented and

Error message

functionality for `nan_as_null` has not been implemented and the parameter will be removed in a future version

Use the default `nan_as_null=False`.

What it means

Raised by PolarsDataFrame.__dataframe__ when nan_as_null=True is passed. The parameter is declared for spec compatibility but its semantics (treating float NaN as null in the interchange layer) was never implemented; only the default False is accepted, and the parameter is slated for removal.

Source

Thrown at py-polars/src/polars/interchange/dataframe.py:64

        ----------
        nan_as_null
            Overwrite null values in the data with `NaN`.

            .. warning::
                This functionality has not been implemented and the parameter will be
                removed in a future version.
                Setting this to `True` will raise a `NotImplementedError`.
        allow_copy
            Allow memory to be copied to perform the conversion. If set to `False`,
            causes conversions that are not zero-copy to fail.
        """
        if nan_as_null:
            msg = (
                "functionality for `nan_as_null` has not been implemented and the"
                " parameter will be removed in a future version"
                "\n\nUse the default `nan_as_null=False`."
            )
            raise NotImplementedError(msg)
        return PolarsDataFrame(self._df, allow_copy=allow_copy)

    @property
    def metadata(self) -> dict[str, Any]:
        """The metadata for the dataframe."""
        return {}

    def num_columns(self) -> int:
        """Return the number of columns in the dataframe."""
        return self._df.width

    def num_rows(self) -> int:
        """Return the number of rows in the dataframe."""
        return self._df.height

    def num_chunks(self) -> int:
        """
        Return the number of chunks the dataframe consists of.

View on GitHub (pinned to df599052da)

Solutions

1. Call with nan_as_null=False (or omit it)
2. If NaN-as-null semantics are needed, normalize on the polars side first: df.with_columns(pl.col(c).fill_nan(None) for float columns)
3. Upgrade the consuming library to a version that no longer passes nan_as_null=True

Example fix

// before
df.__dataframe__(nan_as_null=True)
// after
df.__dataframe__(nan_as_null=False)
# NaN-as-null handled explicitly:
df = df.with_columns(pl.col(pl.Float64).fill_nan(None))

if nan_as_null:
    df = df.with_columns(pl.col(pl.Float64).fill_nan(None))
    nan_as_null = False  # then call __dataframe__(nan_as_null=False)

try:
    dfi = df.__dataframe__(nan_as_null=flag)
except NotImplementedError:
    dfi = df.__dataframe__(nan_as_null=False)

When it happens

Trigger: Calling df.__dataframe__(nan_as_null=True); interchange consumers written against older dataframe-exchange drafts that default or pass nan_as_null=True; copy-pasted examples from legacy interchange tutorials.

Common situations: Upgrading libraries whose interchange integration predates the parameter's deprecation; consumers wanting NaN-as-null semantics for float columns; version drift between a consumer library and a newer polars.

Related errors

• bitmask must be constructed
• functionality for `nan_as_null` has not been implemented and
• non-contiguous buffer must be made contiguous
• __dlpack__
• `describe_categorical` only works on categorical columns

AI-assisted analysis of pola-rs/polars@df599052da (2026-08-16). Data as JSON: /api/errors/fbbd3b589c7fd2a4. Report an issue: GitHub.