You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by "xinrong-meng (via GitHub)" <gi...@apache.org> on 2023/05/11 00:23:11 UTC
[GitHub] [spark] xinrong-meng opened a new pull request, #41125: [SPARK-43440][PYTHON][CONNECT] Support registration of an Arrow-optimized Python UDF
xinrong-meng opened a new pull request, #41125:
URL: https://github.com/apache/spark/pull/41125
### What changes were proposed in this pull request?
The PR proposes to provide support for the registration of an Arrow-optimized Python UDF in both vanilla PySpark and Spark Connect.
### Why are the changes needed?
Currently, when users register an Arrow-optimized Python UDF, it will be registered as a pickled Python UDF and thus, executed without Arrow optimization.
We should support Arrow-optimized Python UDFs registration and execute them with Arrow optimization.
### Does this PR introduce _any_ user-facing change?
Yes.
Previously, a registered Arrow-optimized Python UDF will be executed without Arrow optimization.
Now, it will be executed with Arrow optimization. Enabling or disabling Arrow optimization can produce result differences in some cases.
### How was this patch tested?
Unit test.
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] xinrong-meng commented on a diff in pull request #41125: [SPARK-43440][PYTHON][CONNECT] Support registration of an Arrow-optimized Python UDF
Posted by "xinrong-meng (via GitHub)" <gi...@apache.org>.
xinrong-meng commented on code in PR #41125:
URL: https://github.com/apache/spark/pull/41125#discussion_r1191560291
##########
python/pyspark/sql/udf.py:
##########
@@ -623,25 +623,31 @@ def register(
f = cast("UserDefinedFunctionLike", f)
if f.evalType not in [
PythonEvalType.SQL_BATCHED_UDF,
+ PythonEvalType.SQL_ARROW_BATCHED_UDF,
PythonEvalType.SQL_SCALAR_PANDAS_UDF,
PythonEvalType.SQL_SCALAR_PANDAS_ITER_UDF,
PythonEvalType.SQL_GROUPED_AGG_PANDAS_UDF,
]:
raise PySparkTypeError(
error_class="INVALID_UDF_EVAL_TYPE",
message_parameters={
- "eval_type": "SQL_BATCHED_UDF, SQL_SCALAR_PANDAS_UDF, "
- "SQL_SCALAR_PANDAS_ITER_UDF or SQL_GROUPED_AGG_PANDAS_UDF"
+ "eval_type": "SQL_BATCHED_UDF, SQL_ARROW_BATCHED_UDF, "
+ "SQL_SCALAR_PANDAS_UDF, SQL_SCALAR_PANDAS_ITER_UDF or "
+ "SQL_GROUPED_AGG_PANDAS_UDF"
},
)
- register_udf = _create_udf(
+ source_udf = _create_udf(
f.func,
returnType=f.returnType,
name=name,
evalType=f.evalType,
deterministic=f.deterministic,
- )._unwrapped # type: ignore[attr-defined]
- return_udf = f
+ )
+ if f.evalType == PythonEvalType.SQL_ARROW_BATCHED_UDF:
+ register_udf = _create_arrow_py_udf(source_udf)._unwrapped
Review Comment:
That line is critical.
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] HyukjinKwon closed pull request #41125: [SPARK-43440][PYTHON][CONNECT] Support registration of an Arrow-optimized Python UDF
Posted by "HyukjinKwon (via GitHub)" <gi...@apache.org>.
HyukjinKwon closed pull request #41125: [SPARK-43440][PYTHON][CONNECT] Support registration of an Arrow-optimized Python UDF
URL: https://github.com/apache/spark/pull/41125
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] HyukjinKwon commented on pull request #41125: [SPARK-43440][PYTHON][CONNECT] Support registration of an Arrow-optimized Python UDF
Posted by "HyukjinKwon (via GitHub)" <gi...@apache.org>.
HyukjinKwon commented on PR #41125:
URL: https://github.com/apache/spark/pull/41125#issuecomment-1547145558
Merged to master.
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] xinrong-meng commented on a diff in pull request #41125: [SPARK-43440][PYTHON][CONNECT] Support registration of an Arrow-optimized Python UDF
Posted by "xinrong-meng (via GitHub)" <gi...@apache.org>.
xinrong-meng commented on code in PR #41125:
URL: https://github.com/apache/spark/pull/41125#discussion_r1191560291
##########
python/pyspark/sql/udf.py:
##########
@@ -623,25 +623,31 @@ def register(
f = cast("UserDefinedFunctionLike", f)
if f.evalType not in [
PythonEvalType.SQL_BATCHED_UDF,
+ PythonEvalType.SQL_ARROW_BATCHED_UDF,
PythonEvalType.SQL_SCALAR_PANDAS_UDF,
PythonEvalType.SQL_SCALAR_PANDAS_ITER_UDF,
PythonEvalType.SQL_GROUPED_AGG_PANDAS_UDF,
]:
raise PySparkTypeError(
error_class="INVALID_UDF_EVAL_TYPE",
message_parameters={
- "eval_type": "SQL_BATCHED_UDF, SQL_SCALAR_PANDAS_UDF, "
- "SQL_SCALAR_PANDAS_ITER_UDF or SQL_GROUPED_AGG_PANDAS_UDF"
+ "eval_type": "SQL_BATCHED_UDF, SQL_ARROW_BATCHED_UDF, "
+ "SQL_SCALAR_PANDAS_UDF, SQL_SCALAR_PANDAS_ITER_UDF or "
+ "SQL_GROUPED_AGG_PANDAS_UDF"
},
)
- register_udf = _create_udf(
+ source_udf = _create_udf(
f.func,
returnType=f.returnType,
name=name,
evalType=f.evalType,
deterministic=f.deterministic,
- )._unwrapped # type: ignore[attr-defined]
- return_udf = f
+ )
+ if f.evalType == PythonEvalType.SQL_ARROW_BATCHED_UDF:
+ register_udf = _create_arrow_py_udf(source_udf)._unwrapped
Review Comment:
That's the magic line :).
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org