python/pyspark/sql/pandas/conversion.py - master - githubmirror/spark

python/pyspark/sql/pandas/conversion.py
1 163 строки44 KB
Yicong Huang
[SPARK-46776][PYTHON][FOLLOWUP] Cast large_string/large_binary to the requested type on pyarrow < 19
12 июн 2026, 21:59
12 июн 2026, 21:593b5cbeb
#
# Licensed to the Apache Software Foundation (ASF) under one or more
# contributor license agreements.  See the NOTICE file distributed with
# this work for additional information regarding copyright ownership.
# The ASF licenses this file to You under the Apache License, Version 2.0
# (the "License"); you may not use this file except in compliance with
# the License.  You may obtain a copy of the License at
#
#    http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
import sys
from typing import (
    Any,
    Callable,
    Iterable,
    List,
    Optional,
    Sequence,
    Tuple,
    Union,
    cast,
    no_type_check,
    overload,
    TYPE_CHECKING,
)
from warnings import warn

from pyspark.errors.exceptions.captured import unwrap_spark_exception
from pyspark.util import _load_from_socket
from pyspark.sql.pandas.serializers import ArrowCollectSerializer
from pyspark.sql.pandas.types import _dedup_names
from pyspark.sql.types import (
    ArrayType,
    MapType,
    TimestampType,
    StructType,
    _has_type,
    DataType,
    _create_row,
    StringType,
)
from pyspark.traceback_utils import SCCallSiteSync
from pyspark.errors import PySparkTypeError, PySparkValueError

if TYPE_CHECKING:
    import numpy as np
    import pandas as pd
    import pyarrow as pa
    from py4j.java_gateway import JavaObject

    from pyspark.sql.pandas._typing import DataFrameLike as PandasDataFrameLike
    from pyspark.sql import DataFrame


def create_arrow_array_from_pandas(
    series: "pd.Series",
    spark_type: Optional[DataType],
    *,
    timezone: Optional[str] = None,
    safecheck: bool = False,
    prefers_large_types: bool = False,
) -> "pa.Array":
    """
    Create an Arrow Array from the given pandas.Series and Spark type.

    Parameters
    ----------
    series : pandas.Series
        A single series
    spark_type : DataType, optional
        The Spark return type. If None, pyarrow's inferred type will be used.
    timezone : str, optional
        The timezone to use for timestamp conversions.
    safecheck : bool, optional
        Whether to enable safe type checking during conversion.
    prefers_large_types : bool, optional
        Whether to prefer large Arrow types (e.g., large_string instead of string).

    Returns
    -------
    pyarrow.Array
    """
    import pyarrow as pa
    import pandas as pd
    from pyspark.loose_version import LooseVersion
    from pyspark.sql.pandas.types import to_arrow_type, _create_converter_from_pandas

    if isinstance(series.dtype, pd.CategoricalDtype):
        series = series.astype(series.dtype.categories.dtype)

    # Derive arrow_type from spark_type
    arrow_type = (
        to_arrow_type(spark_type, timezone=timezone, prefers_large_types=prefers_large_types)
        if spark_type is not None
        else None
    )

    if spark_type is not None:
        conv = _create_converter_from_pandas(
            spark_type,
            timezone=timezone,
            error_on_duplicated_field_names=False,
        )
        series = conv(series)

    if hasattr(series.array, "__arrow_array__"):
        mask = None
    else:
        mask = series.isnull()
    try:
        result = pa.Array.from_pandas(series, mask=mask, type=arrow_type, safe=safecheck)
        # SPARK-46776: pyarrow < 19.0.0 ignores the requested ``type`` in the
        # ``__arrow_array__`` protocol used by pyarrow-backed extension dtypes, so a
        # ``string[pyarrow]`` series (backed by ``large_string`` since pandas 2.2.0) can
        # come back as ``large_string`` even when ``string`` was requested, silently
        # corrupting data on the JVM side. Cast back only for this exact (large_)string /
        # (large_)binary offset-width mismatch; pyarrow >= 19.0.0 already honors the type.
        if (
            arrow_type is not None
            and LooseVersion(pa.__version__) < LooseVersion("19.0.0")
            and (
                (pa.types.is_large_string(result.type) and pa.types.is_string(arrow_type))
                or (pa.types.is_large_binary(result.type) and pa.types.is_binary(arrow_type))
            )
        ):
            result = result.cast(arrow_type)
        return result
    except TypeError as e:
        error_msg = (
            "Exception thrown when converting pandas.Series (%s) "
            "with name '%s' to Arrow Array (%s)."
        )
        raise PySparkTypeError(error_msg % (series.dtype, series.name, arrow_type)) from e
    except ValueError as e:
        error_msg = (
            "Exception thrown when converting pandas.Series (%s) "
            "with name '%s' to Arrow Array (%s)."
        )
        if safecheck:
            error_msg = error_msg + (
                " It can be caused by overflows or other "
                "unsafe conversions warned by Arrow. Arrow safe type check "
                "can be disabled by using SQL config "
                "`spark.sql.execution.pandas.convertToArrowArraySafely`."
            )
        raise PySparkValueError(error_msg % (series.dtype, series.name, arrow_type)) from e


def create_arrow_table_from_pandas(
    series_with_types: Iterable[Tuple["pd.Series", Optional[DataType]]],
    *,
    timezone: Optional[str] = None,
    safecheck: bool = False,
    prefers_large_types: bool = False,
) -> "pa.Table":
    """
    Create an Arrow ``Table`` from the given iterable of (series, spark_type) tuples.

    A ``pa.Table`` is used (rather than a single ``pa.RecordBatch``) because
    ``pa.Array.from_pandas`` may return a ``pa.ChunkedArray`` when the input
    pandas Series is backed by a chunked Arrow array (e.g. pyarrow-backed
    extension dtypes such as ``string[pyarrow]``) or when the data exceeds
    the maximum size of a single Arrow array (e.g. string data larger than
    2 GB). ``pa.RecordBatch.from_arrays`` does not accept ``ChunkedArray``,
    but ``pa.Table.from_arrays`` does. Call ``.to_batches()`` on the result
    to obtain a zero-copy list of ``pa.RecordBatch`` aligned on a common
    chunk boundary.

    Parameters
    ----------
    series_with_types : iterable
        Iterable of (series, spark_type) tuples.
    timezone : str, optional
        The timezone to use for timestamp conversions.
    safecheck : bool, optional
        Whether to enable safe type checking during conversion.
    prefers_large_types : bool, optional
        Whether to prefer large Arrow types (e.g., large_string instead of string).

    Returns
    -------
    pyarrow.Table
    """
    import pyarrow as pa

    arrs = [
        create_arrow_array_from_pandas(
            s,
            spark_type,
            timezone=timezone,
            safecheck=safecheck,
            prefers_large_types=prefers_large_types,
        )
        for s, spark_type in series_with_types
    ]
    return pa.Table.from_arrays(arrs, names=["_%d" % i for i in range(len(arrs))])


def _convert_arrow_table_to_pandas(
    arrow_table: "pa.Table",
    schema: "StructType",
    *,
    timezone: Optional[str] = None,
    struct_handling_mode: Optional[str] = None,
    date_as_object: bool = False,
    self_destruct: bool = False,
) -> "PandasDataFrameLike":
    """
    Helper function to convert Arrow table columns to a pandas DataFrame.

    This function applies Spark-specific type converters to Arrow columns and concatenates
    them into a pandas DataFrame.

    Parameters
    ----------
    arrow_table : pyarrow.Table
        The Arrow table to convert
    schema : StructType
        The schema of the DataFrame
    timezone : str or None
        The timezone to use for timestamp conversions (can be None if not configured)
    struct_handling_mode : str or None
        How to handle struct types in pandas ("dict", "row", or "legacy", can be None
         if not configured). If "legacy", it will be converted to "dict" and error checking
         for duplicated field names will be enabled when StructType fields are present.
    date_as_object : bool
        Whether to convert date values to Python datetime.date objects (default: False)
    self_destruct : bool
        Whether to enable memory-efficient self-destruct mode for large tables (default: False)

    Returns
    -------
    pandas.DataFrame
        The converted pandas DataFrame
    """
    import pandas as pd
    from pyspark.sql.pandas.types import _create_converter_to_pandas

    # Build pandas options
    # Pandas DataFrame created from PyArrow uses datetime64[ns] for date type
    # values, but we should use datetime.date to match the behavior with when
    # Arrow optimization is disabled.
    pandas_options = {"coerce_temporal_nanoseconds": True}
    if date_as_object:
        pandas_options["date_as_object"] = True

    # Handle empty columns case
    if len(schema.fields) == 0:
        return arrow_table.to_pandas(**pandas_options)

    # Rename columns to avoid duplicated column names during processing
    temp_col_names = [f"col_{i}" for i in range(len(schema.names))]
    arrow_table = arrow_table.rename_columns(temp_col_names)

    # Configure self-destruct mode for memory efficiency
    if self_destruct and arrow_table.num_rows > 0:
        # Configure PyArrow to use as little memory as possible:
        # self_destruct - free columns as they are converted
        # split_blocks - create a separate Pandas block for each column
        # use_threads - convert one column at a time
        pandas_options.update(
            {
                "self_destruct": True,
                "split_blocks": True,
                "use_threads": False,
            }
        )

    # Handle legacy struct handling mode
    error_on_duplicated_field_names = False
    if struct_handling_mode == "legacy" and any(
        _has_type(f.dataType, StructType) for f in schema.fields
    ):
        error_on_duplicated_field_names = True
        struct_handling_mode = "dict"

    # Convert arrow columns to pandas Series
    column_data = (arrow_col.to_pandas(**pandas_options) for arrow_col in arrow_table.columns)

    # Apply Spark-specific type converters to each column
    pdf = pd.concat(
        objs=cast(
            Sequence[pd.Series],
            (
                _create_converter_to_pandas(
                    field.dataType,
                    field.nullable,
                    timezone=timezone,
                    struct_in_pandas=struct_handling_mode,
                    error_on_duplicated_field_names=error_on_duplicated_field_names,
                )(series)
                for series, field in zip(column_data, schema.fields)
            ),
        ),
        axis="columns",
    )

    # Restore original column names (including duplicates)
    pdf.columns = pd.Index(schema.names)

    return pdf


class PandasConversionMixin:
    """
    Mix-in for the conversion from Spark to pandas and PyArrow. Currently, only
    :class:`DataFrame` can use this class.
    """

    def _to_pandas(self, **kwargs: Any) -> "PandasDataFrameLike":
        from pyspark.sql.dataframe import DataFrame

        assert isinstance(self, DataFrame)

        from pyspark.sql.pandas.types import _create_converter_to_pandas
        from pyspark.sql.pandas.utils import require_minimum_pandas_version

        require_minimum_pandas_version()

        import pandas as pd

        (
            sessionLocalTimeZone,
            arrowPySparkEnabled,
            arrowUseLargeVarTypes,
            arrowPySparkFallbackEnabled,
            arrowPySparkSelfDestructEnabled,
            pandasStructHandlingMode,
        ) = self.sparkSession._jconf.getConfs(
            [
                "spark.sql.session.timeZone",
                "spark.sql.execution.arrow.pyspark.enabled",
                "spark.sql.execution.arrow.useLargeVarTypes",
                "spark.sql.execution.arrow.pyspark.fallback.enabled",
                "spark.sql.execution.arrow.pyspark.selfDestruct.enabled",
                "spark.sql.execution.pandas.structHandlingMode",
            ]
        )

        # if pandasStructHandlingMode is explicitly set, override the runtime config
        if "pandasStructHandlingMode" in kwargs:
            pandasStructHandlingMode = str(kwargs["pandasStructHandlingMode"])

        prefers_large_var_types = arrowUseLargeVarTypes == "true"

        if arrowPySparkEnabled == "true":
            use_arrow = True
            try:
                from pyspark.sql.pandas.types import to_arrow_schema
                from pyspark.sql.pandas.utils import require_minimum_pyarrow_version

                require_minimum_pyarrow_version()
                arrow_schema = to_arrow_schema(
                    self.schema,
                    timezone="UTC",
                    prefers_large_types=prefers_large_var_types,
                )
            except Exception as e:
                if arrowPySparkFallbackEnabled == "true":
                    msg = (
                        "toPandas attempted Arrow optimization because "
                        "'spark.sql.execution.arrow.pyspark.enabled' is set to true; however, "
                        "failed by the reason below:\n  %s\n"
                        "Attempting non-optimization as "
                        "'spark.sql.execution.arrow.pyspark.fallback.enabled' is set to "
                        "true." % str(e)
                    )
                    warn(msg)
                    use_arrow = False
                else:
                    msg = (
                        "toPandas attempted Arrow optimization because "
                        "'spark.sql.execution.arrow.pyspark.enabled' is set to true, but has "
                        "reached the error below and will not continue because automatic fallback "
                        "with 'spark.sql.execution.arrow.pyspark.fallback.enabled' has been set to "
                        "false.\n  %s" % str(e)
                    )
                    warn(msg)
                    raise

            # Try to use Arrow optimization when the schema is supported and the required version
            # of PyArrow is found, if 'spark.sql.execution.arrow.pyspark.enabled' is enabled.
            if use_arrow:
                try:
                    import pyarrow as pa

                    batches = self._collect_as_arrow(
                        split_batches=arrowPySparkSelfDestructEnabled == "true",
                        prefers_large_var_types=prefers_large_var_types,
                    )

                    if len(batches) > 0:
                        table = pa.Table.from_batches(batches)
                    else:
                        # empty dataset
                        table = arrow_schema.empty_table()

                    # Ensure only the table has a reference to the batches, so that
                    # self_destruct (if enabled) is effective
                    del batches

                    pdf = _convert_arrow_table_to_pandas(
                        arrow_table=table,
                        schema=self.schema,
                        timezone=sessionLocalTimeZone,
                        struct_handling_mode=pandasStructHandlingMode,
                        date_as_object=True,
                        self_destruct=arrowPySparkSelfDestructEnabled == "true",
                    )

                    return pdf

                except Exception as e:
                    # We might have to allow fallback here as well but multiple Spark jobs can
                    # be executed. So, simply fail in this case for now.
                    msg = (
                        "toPandas attempted Arrow optimization because "
                        "'spark.sql.execution.arrow.pyspark.enabled' is set to true, but has "
                        "reached the error below and can not continue. Note that "
                        "'spark.sql.execution.arrow.pyspark.fallback.enabled' does not have an "
                        "effect on failures in the middle of "
                        "computation.\n  %s" % str(e)
                    )
                    warn(msg)
                    raise

        # Below is toPandas without Arrow optimization.
        rows = self.collect()
        if len(rows) > 0:
            pdf = pd.DataFrame.from_records(rows, index=range(len(rows)), columns=self.columns)
        else:
            pdf = pd.DataFrame(columns=self.columns)

        if len(pdf.columns) > 0:
            timezone = sessionLocalTimeZone
            struct_in_pandas = pandasStructHandlingMode

            return pd.concat(
                [
                    _create_converter_to_pandas(
                        field.dataType,
                        field.nullable,
                        timezone=timezone,
                        struct_in_pandas=(
                            "row" if struct_in_pandas == "legacy" else struct_in_pandas
                        ),
                        error_on_duplicated_field_names=False,
                        timestamp_utc_localized=False,
                    )(pser)
                    for (_, pser), field in zip(pdf.items(), self.schema.fields)
                ],
                axis="columns",
            )
        else:
            return pdf

    def toArrow(self) -> "pa.Table":
        from pyspark.sql.dataframe import DataFrame

        assert isinstance(self, DataFrame)

        from pyspark.sql.pandas.types import to_arrow_schema
        from pyspark.sql.pandas.utils import require_minimum_pyarrow_version

        require_minimum_pyarrow_version()

        (
            arrowUseLargeVarTypes,
            arrowPySparkSelfDestructEnabled,
        ) = self.sparkSession._jconf.getConfs(
            [
                "spark.sql.execution.arrow.useLargeVarTypes",
                "spark.sql.execution.arrow.pyspark.selfDestruct.enabled",
            ]
        )

        prefers_large_var_types = arrowUseLargeVarTypes == "true"
        schema = to_arrow_schema(
            self.schema,
            error_on_duplicated_field_names_in_struct=True,
            timezone="UTC",
            prefers_large_types=prefers_large_var_types,
        )

        import pyarrow as pa

        self_destruct = arrowPySparkSelfDestructEnabled == "true"
        batches = self._collect_as_arrow(
            split_batches=self_destruct,
            empty_list_if_zero_records=False,
            prefers_large_var_types=prefers_large_var_types,
        )
        table = pa.Table.from_batches(batches).cast(schema)
        # Ensure only the table has a reference to the batches, so that
        # self_destruct (if enabled) is effective
        del batches
        return table

    def _collect_as_arrow(
        self,
        split_batches: bool = False,
        empty_list_if_zero_records: bool = True,
        prefers_large_var_types: bool = False,
    ) -> List["pa.RecordBatch"]:
        """
        Returns all records as a list of Arrow RecordBatches. PyArrow must be installed
        and available on driver and worker Python environments.
        This is an experimental feature.

        :param split_batches: split batches such that each column is in its own allocation, so
            that the selfDestruct optimization is effective; default False.

        :param empty_list_if_zero_records: If True (the default), returns an empty list if the
            result has 0 records. Otherwise, returns a list of length 1 containing an empty
            Arrow RecordBatch which includes the schema.

        .. note:: Experimental.
        """
        from pyspark.sql.dataframe import DataFrame

        assert isinstance(self, DataFrame)

        with SCCallSiteSync(self._sc):
            (
                port,
                auth_secret,
                jsocket_auth_server,
            ) = self._jdf.collectAsArrowToPython()

        # Collect list of un-ordered batches where last element is a list of correct order indices
        try:
            with _load_from_socket((port, auth_secret), ArrowCollectSerializer()) as batch_stream:
                if split_batches:
                    # When spark.sql.execution.arrow.pyspark.selfDestruct.enabled, ensure
                    # each column in each record batch is contained in its own allocation.
                    # Otherwise, selfDestruct does nothing; it frees each column as its
                    # converted, but each column will actually be a list of slices of record
                    # batches, and so no memory is actually freed until all columns are
                    # converted.
                    import pyarrow as pa

                    results = []
                    for batch_or_indices in batch_stream:
                        if isinstance(batch_or_indices, pa.RecordBatch):
                            batch_or_indices = pa.RecordBatch.from_arrays(
                                [
                                    # This call actually reallocates the array
                                    pa.concat_arrays([array])
                                    for array in batch_or_indices
                                ],
                                schema=batch_or_indices.schema,
                            )
                        results.append(batch_or_indices)
                else:
                    results = list(batch_stream)
        finally:
            with unwrap_spark_exception():
                # Join serving thread and raise any exceptions from collectAsArrowToPython
                jsocket_auth_server.getResult()

        # Separate RecordBatches from batch order indices in results
        batches = results[:-1]
        batch_order = results[-1]

        if len(batches) or empty_list_if_zero_records:
            # Re-order the batch list using the correct order
            return [batches[i] for i in batch_order]
        else:
            from pyspark.sql.pandas.types import to_arrow_schema
            import pyarrow as pa

            schema = to_arrow_schema(
                self.schema, timezone="UTC", prefers_large_types=prefers_large_var_types
            )
            empty_arrays = [pa.array([], type=field.type) for field in schema]
            return [pa.RecordBatch.from_arrays(empty_arrays, schema=schema)]


class SparkConversionMixin:
    """
    Min-in for the conversion from pandas and PyArrow to Spark. Currently, only
    :class:`SparkSession` can use this class.
    """

    _jsparkSession: "JavaObject"

    @overload
    def createDataFrame(
        self, data: "PandasDataFrameLike", samplingRatio: Optional[float] = ...
    ) -> "DataFrame": ...

    @overload
    def createDataFrame(
        self, data: "pa.Table", samplingRatio: Optional[float] = ...
    ) -> "DataFrame": ...

    @overload
    def createDataFrame(
        self,
        data: "PandasDataFrameLike",
        schema: Union[StructType, str],
        verifySchema: bool = ...,
    ) -> "DataFrame": ...

    @overload
    def createDataFrame(
        self,
        data: "pa.Table",
        schema: Union[StructType, str],
        verifySchema: bool = ...,
    ) -> "DataFrame": ...

    def createDataFrame(  # type: ignore[misc]
        self,
        data: Union["PandasDataFrameLike", "pa.Table"],
        schema: Optional[Union[StructType, List[str]]] = None,
        samplingRatio: Optional[float] = None,
        verifySchema: bool = True,
    ) -> "DataFrame":
        from pyspark.sql import SparkSession

        assert isinstance(self, SparkSession)

        (
            timestampType,
            sessionLocalTimeZone,
            arrowPySparkEnabled,
            arrowUseLargeVarTypes,
            arrowPySparkFallbackEnabled,
            arrowMaxRecordsPerBatch,
            arrowSafeTypeConversion,
            inferPandasDictAsMap,
        ) = self._jconf.getConfs(
            [
                "spark.sql.timestampType",
                "spark.sql.session.timeZone",
                "spark.sql.execution.arrow.pyspark.enabled",
                "spark.sql.execution.arrow.useLargeVarTypes",
                "spark.sql.execution.arrow.pyspark.fallback.enabled",
                "spark.sql.execution.arrow.maxRecordsPerBatch",
                "spark.sql.execution.pandas.convertToArrowArraySafely",
                "spark.sql.execution.pandas.inferPandasDictAsMap",
            ]
        )

        prefer_timestamp_ntz = timestampType == "TIMESTAMP_NTZ"
        prefers_large_var_types = arrowUseLargeVarTypes == "true"
        timezone = sessionLocalTimeZone
        arrow_batch_size = int(arrowMaxRecordsPerBatch)
        selfcheck = arrowSafeTypeConversion == "true"
        infer_pandas_dict_as_map = inferPandasDictAsMap == "true"

        if type(data).__name__ == "Table":
            # `data` is a PyArrow Table
            from pyspark.sql.pandas.utils import require_minimum_pyarrow_version

            require_minimum_pyarrow_version()

            import pyarrow as pa

            assert isinstance(data, pa.Table)

            # If no schema supplied by user then get the names of columns only
            if schema is None:
                schema = data.schema.names

            return self._create_from_arrow_table(
                data,
                schema,
                timezone,
                prefer_timestamp_ntz,
                prefers_large_var_types,
                arrow_batch_size,
            )

        # `data` is a PandasDataFrameLike object
        from pyspark.sql.pandas.utils import require_minimum_pandas_version

        require_minimum_pandas_version()

        # If no schema supplied by user then get the names of columns only
        if schema is None:
            schema = [str(x) if not isinstance(x, str) else x for x in data.columns]

        if arrowPySparkEnabled == "true" and len(data) > 0:
            try:
                return self._create_from_pandas_with_arrow(
                    data,
                    schema,
                    timezone,
                    prefer_timestamp_ntz,
                    prefers_large_var_types,
                    arrow_batch_size,
                    selfcheck,
                    infer_pandas_dict_as_map,
                )
            except Exception as e:
                if arrowPySparkFallbackEnabled == "true":
                    msg = (
                        "createDataFrame attempted Arrow optimization because "
                        "'spark.sql.execution.arrow.pyspark.enabled' is set to true; however, "
                        "failed by the reason below:\n  %s\n"
                        "Attempting non-optimization as "
                        "'spark.sql.execution.arrow.pyspark.fallback.enabled' is set to "
                        "true." % str(e)
                    )
                    warn(msg)
                else:
                    msg = (
                        "createDataFrame attempted Arrow optimization because "
                        "'spark.sql.execution.arrow.pyspark.enabled' is set to true, but has "
                        "reached the error below and will not continue because automatic "
                        "fallback with 'spark.sql.execution.arrow.pyspark.fallback.enabled' "
                        "has been set to false.\n  %s" % str(e)
                    )
                    warn(msg)
                    raise
        converted_data = self._convert_from_pandas(data, schema, timezone, prefer_timestamp_ntz)
        return self._create_dataframe(converted_data, schema, samplingRatio, verifySchema)

    def _convert_from_pandas(
        self,
        pdf: "PandasDataFrameLike",
        schema: Union[StructType, str, List[str]],
        timezone: str,
        prefer_timestamp_ntz: bool,
    ) -> List:
        """
        Convert a pandas.DataFrame to list of records that can be used to make a DataFrame

        Returns
        -------
        list
            list of records
        """
        from pyspark.sql import SparkSession

        assert isinstance(self, SparkSession)

        if timezone is not None:
            from pyspark.sql.pandas.types import (
                _check_series_convert_timestamps_tz_local,
                _get_local_timezone,
            )
            import pandas as pd
            from pandas.core.dtypes.common import is_timedelta64_dtype

            copied = False
            if isinstance(schema, StructType):

                def _create_converter(data_type: DataType) -> Callable[[pd.Series], pd.Series]:
                    if isinstance(data_type, TimestampType):

                        def correct_timestamp(pser: pd.Series) -> pd.Series:
                            return _check_series_convert_timestamps_tz_local(pser, timezone)

                        return correct_timestamp

                    def _converter(dt: DataType) -> Optional[Callable[[Any], Any]]:
                        if isinstance(dt, ArrayType):
                            element_conv = _converter(dt.elementType) or (lambda x: x)

                            def convert_array(value: Any) -> Any:
                                if value is None:
                                    return None
                                else:
                                    return [element_conv(v) for v in value]

                            return convert_array

                        elif isinstance(dt, MapType):
                            key_conv = _converter(dt.keyType) or (lambda x: x)
                            value_conv = _converter(dt.valueType) or (lambda x: x)

                            def convert_map(value: Any) -> Any:
                                if value is None:
                                    return None
                                else:
                                    return {key_conv(k): value_conv(v) for k, v in value.items()}

                            return convert_map

                        elif isinstance(dt, StructType):
                            field_names = dt.names
                            dedup_field_names = _dedup_names(field_names)
                            field_convs = [
                                _converter(f.dataType) or (lambda x: x) for f in dt.fields
                            ]

                            def convert_struct(value: Any) -> Any:
                                if value is None:
                                    return None
                                elif isinstance(value, dict):
                                    _values = [
                                        field_convs[i](value.get(name, None))
                                        for i, name in enumerate(dedup_field_names)
                                    ]
                                    return _create_row(field_names, _values)
                                else:
                                    _values = [
                                        field_convs[i](value[i]) for i, name in enumerate(value)
                                    ]
                                    return _create_row(field_names, _values)

                            return convert_struct

                        elif isinstance(dt, TimestampType):

                            def convert_timestamp(value: Any) -> Any:
                                if value is None:
                                    return None
                                else:
                                    return (
                                        pd.Timestamp(value)
                                        .tz_localize(timezone, ambiguous=False)
                                        .tz_convert(_get_local_timezone())
                                        .tz_localize(None)
                                        .to_pydatetime()
                                    )

                            return convert_timestamp

                        else:
                            return None

                    conv = _converter(data_type)
                    if conv is not None:
                        return lambda pser: pser.apply(conv)
                    else:
                        return lambda pser: pser

                if len(pdf.columns) > 0:
                    pdf = pd.concat(
                        [
                            _create_converter(field.dataType)(pser)
                            for (_, pser), field in zip(pdf.items(), schema.fields)
                        ],
                        axis="columns",
                    )
                    copied = True
            else:
                should_localize = not prefer_timestamp_ntz
                for column, series in pdf.items():
                    s = series
                    if (
                        should_localize
                        and isinstance(s.dtype, pd.DatetimeTZDtype)
                        and s.dt.tz is not None
                    ):
                        s = _check_series_convert_timestamps_tz_local(series, timezone)
                    if s is not series:
                        if not copied:
                            # Copy once if the series is modified to prevent the original
                            # Pandas DataFrame from being updated
                            pdf = pdf.copy()
                            copied = True
                        pdf[column] = s

            for column, series in pdf.items():
                if is_timedelta64_dtype(series):
                    if not copied:
                        pdf = pdf.copy()
                        copied = True
                    # Explicitly set the timedelta as object so the output of numpy records can
                    # hold the timedelta instances as are. Otherwise, it converts to the internal
                    # numeric values.
                    ser = pdf[column]
                    pdf[column] = pd.Series(
                        ser.dt.to_pytimedelta(), index=ser.index, dtype="object", name=ser.name
                    )

        # Handle the 0-column case separately to preserve row count
        if len(pdf.columns) == 0:
            from pyspark.sql import Row

            return [Row()] * len(pdf)

        # Convert pandas.DataFrame to list of numpy records
        np_records = pdf.set_axis(
            [f"col_{i}" for i in range(len(pdf.columns))], axis="columns"
        ).to_records(index=False)

        # Check if any columns need to be fixed for Spark to infer properly
        if len(np_records) > 0:
            record_dtype = self._get_numpy_record_dtype(np_records[0])
            if record_dtype is not None:
                return [r.astype(record_dtype).tolist() for r in np_records]

        # Convert list of numpy records to python lists
        return [r.tolist() for r in np_records]

    def _get_numpy_record_dtype(self, rec: "np.recarray") -> Optional["np.dtype"]:
        """
        Used when converting a pandas.DataFrame to Spark using to_records(), this will correct
        the dtypes of fields in a record so they can be properly loaded into Spark.

        Parameters
        ----------
        rec : numpy.record
            a numpy record to check field dtypes

        Returns
        -------
        numpy.dtype
            corrected dtype for a numpy.record or None if no correction needed
        """
        import numpy as np

        cur_dtypes = rec.dtype
        col_names = cur_dtypes.names
        record_type_list = []
        has_rec_fix = False
        for i in range(len(cur_dtypes)):  # type: ignore[arg-type]
            curr_type = cur_dtypes[i]
            # If type is a datetime64 timestamp, convert to microseconds
            # NOTE: if dtype is datetime[ns] then np.record.tolist() will output values as longs,
            # conversion from [us] or lower will lead to py datetime objects, see SPARK-22417
            if curr_type == np.dtype("datetime64[ns]"):
                curr_type = "datetime64[us]"  # type: ignore[assignment]
                has_rec_fix = True
            record_type_list.append((str(col_names[i]), curr_type))  # type: ignore[index]
        return np.dtype(record_type_list) if has_rec_fix else None

    def _create_from_pandas_with_arrow(
        self,
        pdf: "PandasDataFrameLike",
        schema: Union[StructType, List[str]],
        timezone: str,
        prefer_timestamp_ntz: bool,
        prefers_large_var_types: bool,
        arrow_batch_size: int,
        safecheck: bool,
        infer_pandas_dict_as_map: bool,
    ) -> "DataFrame":
        """
        Create a DataFrame from a given pandas.DataFrame by slicing it into partitions, converting
        to Arrow data, then sending to the JVM to parallelize. If a schema is passed in, the
        data types will be used to coerce the data in Pandas to Arrow conversion.
        """
        from pyspark.sql import SparkSession
        from pyspark.sql.dataframe import DataFrame

        assert isinstance(self, SparkSession)

        from pyspark.sql.pandas.serializers import ArrowStreamSerializer
        from pyspark.sql.types import TimestampType
        from pyspark.sql.pandas.types import (
            from_arrow_type,
            _deduplicate_field_names,
        )
        from pyspark.sql.pandas.utils import (
            require_minimum_pandas_version,
            require_minimum_pyarrow_version,
        )

        require_minimum_pandas_version()
        require_minimum_pyarrow_version()

        import pandas as pd
        from pandas.api.types import is_datetime64_dtype
        import pyarrow as pa

        # Create the Spark schema from list of names passed in with Arrow types
        if isinstance(schema, (list, tuple)):
            arrow_schema = pa.Schema.from_pandas(pdf, preserve_index=False)
            struct = StructType()
            if infer_pandas_dict_as_map:
                spark_type: Union[MapType, DataType]
                for name, field in zip(schema, arrow_schema):
                    field_type = field.type
                    if isinstance(field_type, pa.StructType):
                        if len(field_type) == 0:
                            raise PySparkValueError(
                                errorClass="CANNOT_INFER_EMPTY_SCHEMA",
                                messageParameters={},
                            )
                        arrow_type = field_type.field(0).type
                        spark_type = MapType(
                            StringType(), from_arrow_type(arrow_type, prefer_timestamp_ntz)
                        )
                    else:
                        spark_type = from_arrow_type(field_type, prefer_timestamp_ntz)
                    struct.add(name, spark_type, nullable=field.nullable)
            else:
                for name, field in zip(schema, arrow_schema):
                    struct.add(
                        name,
                        from_arrow_type(field.type, prefer_timestamp_ntz),
                        nullable=field.nullable,
                    )
            schema = struct

        # Determine arrow types to coerce data when creating batches
        if isinstance(schema, StructType):
            spark_types = [_deduplicate_field_names(f.dataType) for f in schema.fields]
        elif isinstance(schema, DataType):
            raise PySparkTypeError(
                errorClass="UNSUPPORTED_DATA_TYPE_FOR_ARROW",
                messageParameters={"data_type": str(schema)},
            )
        else:
            # Any timestamps must be coerced to be compatible with Spark
            spark_types = [
                (
                    TimestampType()
                    if is_datetime64_dtype(t) or isinstance(t, pd.DatetimeTZDtype)
                    else None
                )
                for t in pdf.dtypes
            ]

        # Slice the DataFrame to be batched
        step = arrow_batch_size
        step = step if step > 0 else len(pdf)
        pdf_slices = (pdf.iloc[start : start + step] for start in range(0, len(pdf), step))

        # Handle the 0-column case separately to preserve row count.
        # pa.RecordBatch.from_pandas preserves num_rows via pandas index metadata.
        if len(pdf.columns) == 0:
            arrow_batches = [pa.RecordBatch.from_pandas(pdf_slice) for pdf_slice in pdf_slices]
        else:
            # Each slice may produce more than one RecordBatch when a column is
            # backed by a ChunkedArray, so flatten the per-slice tables.
            arrow_batches = [
                b
                for pdf_slice in pdf_slices
                for b in create_arrow_table_from_pandas(
                    [(c, t) for (_, c), t in zip(pdf_slice.items(), spark_types)],
                    timezone=timezone,
                    safecheck=safecheck,
                    prefers_large_types=prefers_large_var_types,
                ).to_batches()
            ]

        jsparkSession = self._jsparkSession

        ser = ArrowStreamSerializer()

        @no_type_check
        def reader_func(temp_filename):
            return self._jvm.PythonSQLUtils.readArrowStreamFromFile(temp_filename)

        @no_type_check
        def create_iter_server():
            return self._jvm.ArrowIteratorServer()

        # Create Spark DataFrame from Arrow stream file, using one batch per partition
        jiter = self._sc._serialize_to_jvm(arrow_batches, ser, reader_func, create_iter_server)
        assert self._jvm is not None
        jdf = self._jvm.PythonSQLUtils.toDataFrame(jiter, schema.json(), jsparkSession)
        df = DataFrame(jdf, self)
        df._schema = schema
        return df

    def _create_from_arrow_table(
        self,
        table: "pa.Table",
        schema: Union[StructType, List[str]],
        timezone: str,
        prefer_timestamp_ntz: bool,
        prefers_large_var_types: bool,
        arrow_batch_size: int,
    ) -> "DataFrame":
        """
        Create a DataFrame from a given pyarrow.Table by slicing it into partitions then
        sending to the JVM to parallelize.
        """
        from pyspark.sql import SparkSession
        from pyspark.sql.dataframe import DataFrame

        assert isinstance(self, SparkSession)

        from pyspark.sql.pandas.serializers import ArrowStreamSerializer
        from pyspark.sql.pandas.types import (
            from_arrow_type,
            from_arrow_schema,
            to_arrow_schema,
            _check_arrow_table_timestamps_localize,
        )
        from pyspark.sql.pandas.utils import require_minimum_pyarrow_version

        require_minimum_pyarrow_version()

        # Create the Spark schema from list of names passed in with Arrow types
        if isinstance(schema, (list, tuple)):
            table = table.rename_columns(schema)
            arrow_schema = table.schema
            struct = StructType()
            for name, field in zip(schema, arrow_schema):
                struct.add(
                    name,
                    from_arrow_type(field.type, prefer_timestamp_ntz),
                    nullable=field.nullable,
                )
            schema = struct

        if not isinstance(schema, StructType):
            schema = from_arrow_schema(table.schema, prefer_timestamp_ntz=prefer_timestamp_ntz)

        # Skip cast for 0-column tables as it loses row count
        if len(schema.fields) > 0:
            table = _check_arrow_table_timestamps_localize(table, schema, True, timezone).cast(
                to_arrow_schema(
                    schema,
                    error_on_duplicated_field_names_in_struct=True,
                    timezone="UTC",
                    prefers_large_types=prefers_large_var_types,
                )
            )

        # Chunk the Arrow Table into RecordBatches
        chunk_size = arrow_batch_size
        arrow_data = table.to_batches(max_chunksize=chunk_size)

        jsparkSession = self._jsparkSession

        ser = ArrowStreamSerializer()

        @no_type_check
        def reader_func(temp_filename):
            return self._jvm.PythonSQLUtils.readArrowStreamFromFile(temp_filename)

        @no_type_check
        def create_iter_server():
            return self._jvm.ArrowIteratorServer()

        # Create Spark DataFrame from Arrow stream file, using one batch per partition
        jiter = self._sc._serialize_to_jvm(arrow_data, ser, reader_func, create_iter_server)
        assert self._jvm is not None
        jdf = self._jvm.PythonSQLUtils.toDataFrame(jiter, schema.json(), jsparkSession)
        df = DataFrame(jdf, self)
        df._schema = schema
        return df


def _test() -> None:
    import doctest
    from pyspark.sql import SparkSession
    import pyspark.sql.pandas.conversion

    globs = pyspark.sql.pandas.conversion.__dict__.copy()
    spark = (
        SparkSession.builder.master("local[4]").appName("sql.pandas.conversion tests").getOrCreate()
    )
    globs["spark"] = spark
    failure_count, test_count = doctest.testmod(
        pyspark.sql.pandas.conversion,
        globs=globs,
        optionflags=doctest.ELLIPSIS | doctest.NORMALIZE_WHITESPACE | doctest.REPORT_NDIFF,
    )
    spark.stop()
    if failure_count:
        sys.exit(-1)


if __name__ == "__main__":
    _test()