ueshin commented on code in PR #40725:
URL: https://github.com/apache/spark/pull/40725#discussion_r1172953234
##########
python/pyspark/sql/connect/udf.py:
##########
@@ -47,6 +49,47 @@
from pyspark.sql.types import StringType
+def _create_py_udf(
+ f: Callable[..., Any],
+ returnType: "DataTypeOrString",
+ evalType: int,
+ useArrow: Optional[bool] = None,
+) -> "UserDefinedFunctionLike":
+ from pyspark.sql.udf import _create_arrow_py_udf
+ from pyspark.sql.connect.session import _active_spark_session
+
+ if _active_spark_session is None:
+ is_arrow_enabled = False
+ else:
+ is_arrow_enabled = (
+
_active_spark_session.conf.get("spark.sql.execution.pythonUDF.arrow.enabled")
== "true"
+ if useArrow is None
+ else useArrow
+ )
+
+ regular_udf = _create_udf(f, returnType, evalType)
+ return_type = regular_udf.returnType
+ try:
+ is_func_with_args = len(getfullargspec(f).args) > 0
+ except TypeError:
+ is_func_with_args = False
+ is_output_atomic_type = (
+ not isinstance(return_type, StructType)
+ and not isinstance(return_type, MapType)
+ and not isinstance(return_type, ArrayType)
+ )
+ if is_arrow_enabled:
+ if is_output_atomic_type and is_func_with_args:
+ return _create_arrow_py_udf(regular_udf)
+ else:
+ warnings.warn(
+ "Arrow optimization for Python UDFs cannot be enabled.",
+ UserWarning,
Review Comment:
If it's only a warning, we should return some result? Otherwise, should
raise an exception?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
