pyarrow is required for converting a pandas dataframe to…

Error message

pyarrow is required for converting a pandas dataframe to Polars, unless each of its columns is a simple numpy-backed one (e.g. 'int64', 'bool', 'float32' - not 'Int64')

What it means

Raised by pandas_to_pydf when pyarrow is not installed and the pandas frame contains columns whose backing dtypes are not simple numpy types (e.g. pandas nullable 'Int64', string dtype, categoricals). Without pyarrow, polars can only convert plain numpy-backed columns and raises ImportError otherwise.

Solutions

Install pyarrow: pip install pyarrow
Cast extension dtypes to numpy equivalents before conversion, e.g. df = df.astype({'col': 'int64'})
Use pl.from_pandas(df) only after ensuring all columns are numpy-backed, or convert column-by-column to polars Series

Example fix

// before
pl.from_pandas(df)  # df has pandas 'Int64' dtype, no pyarrow installed
// after
# pip install pyarrow
pl.from_pandas(df)

Defensive patterns

Strategy: fallback

Validate before calling

import importlib.util
if importlib.util.find_spec('pyarrow') is None:
    raise ImportError("pyarrow required for this pandas conversion; pip install pyarrow")

Type guard

def pyarrow_available() -> bool:
    import importlib.util
    return importlib.util.find_spec('pyarrow') is not None

Try / catch

try:
    df_pl = pl.from_pandas(df)
except ImportError:
    df_pl = pl.DataFrame({str(c): df[c].to_numpy() for c in df.columns})

Prevention

Install the pandas extra: pip install 'polars[pandas]'
Pin pyarrow in requirements/CI images
Avoid pandas nullable extension dtypes ('Int64', 'boolean') on the hot conversion path, or cast to numpy dtypes first

When it happens

Trigger: pl.from_pandas(df) or pl.DataFrame(df) in an environment without pyarrow, where the frame has extension dtypes such as 'Int64', 'boolean', 'string', or datetime-with-timezone columns.

Common situations: Fresh virtualenv/CI image missing the pyarrow extra (pip install 'polars[pandas]'), pandas code migrated to nullable dtypes ('Int64') that used to convert fine, minimal deployments trimming dependencies.

Understand the failure class

Background: "not installed", "pip install", "required for": how missing-dependency errors surface across open-source libraries — this error's family across 34 libraries.

Related errors

AI-assisted analysis of pola-rs/polars@fe841f959e (2026-09-18). Data as JSON: /api/errors/1e51bb4672a1b8f4. Report an issue: GitHub.

Appendix: source

Thrown at py-polars/src/polars/_utils/construction/dataframe.py:1129

            return PyDataFrame.empty_with_height(data.shape[0])

        if all(is_simple_numpy_backed_pandas_series(data[col]) for col in data.columns):
            # Convert via NumPy directly, no PyArrow needed.
            return pl.DataFrame(
                {str(col): data[col].to_numpy() for col in data.columns},
                schema=schema,
                strict=strict,
                schema_overrides=schema_overrides,
                nan_to_null=nan_to_null,
            )._df

    if not _PYARROW_AVAILABLE:
        msg = (
            "pyarrow is required for converting a pandas dataframe to Polars, "
            "unless each of its columns is a simple numpy-backed one "
            "(e.g. 'int64', 'bool', 'float32' - not 'Int64')"
        )
        raise ImportError(msg)
    arrow_dict = {}
    length = data.shape[0]

    if convert_index:
        for idxcol in data.index.names:
            arrow_dict[str(idxcol)] = plc.pandas_series_to_arrow(
                # get_level_values accepts `int | str`
                # but `index.names` returns `Hashable`
                data.index.get_level_values(idxcol),  # type: ignore[arg-type, unused-ignore]
                nan_to_null=nan_to_null,
                length=length,
            )

    for col_idx, col_data in data.items():
        arrow_dict[str(col_idx)] = plc.pandas_series_to_arrow(
            col_data, nan_to_null=nan_to_null, length=length
        )

View on GitHub (pinned to fe841f959e)