You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@spark.apache.org by yi...@apache.org on 2026/09/01 04:51:11 UTC
(spark) branch master updated: [SPARK-58601][PYTHON][FOLLOWUP] Default mapInBatch acceptAnyIterable to true
This is an automated email from the ASF dual-hosted git repository.
Yicong-Huang pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/spark.git
The following commit(s) were added to refs/heads/master by this push:
new 9d50aec5a28c [SPARK-58601][PYTHON][FOLLOWUP] Default mapInBatch acceptAnyIterable to true
9d50aec5a28c is described below
commit 9d50aec5a28c46dee9317187fd38f30936e33752
Author: Yicong Huang <17...@users.noreply.github.com>
AuthorDate: Tue Sep 1 04:50:53 2026 +0000
[SPARK-58601][PYTHON][FOLLOWUP] Default mapInBatch acceptAnyIterable to true
### What changes were proposed in this pull request?
Follow-up to SPARK-58601, which tightened the `mapInPandas`/`mapInArrow` return-value contract to require a strict `Iterator` by default and added the internal escape-hatch config `spark.sql.execution.pythonUDF.mapInBatch.legacy.acceptAnyIterable.enabled` (default `false`). This PR flips that config's default to `true`, so the pre-4.3.0 lenient behavior (accepting any iterable such as a returned `list`) remains the default and can be disabled by setting the flag to `false` to opt into [...]
### Why are the changes needed?
Requiring a strict `Iterator` by default is a user-facing breaking change: a `mapInPandas`/`mapInArrow` UDF that returns a non-`Iterator` iterable (e.g. a `list`) previously worked and would start raising `UDF_RETURN_TYPE`. Defaulting the escape hatch to `true` preserves the established behavior for existing workloads while keeping the flag available for users who want the stricter contract, deferring any default change to a later decision.
### Does this PR introduce _any_ user-facing change?
Yes, relative to the unreleased SPARK-58601 change on the unreleased branch. With this PR a `mapInPandas`/`mapInArrow` UDF that returns any iterable (e.g. a `list`) is accepted again by default, as it was before SPARK-58601. Setting `spark.sql.execution.pythonUDF.mapInBatch.legacy.acceptAnyIterable.enabled=false` restores the strict `Iterator`-only contract. Compared to released Spark versions there is no behavior change.
### How was this patch tested?
Updated `test_pandas_map.py` and `test_arrow_map.py`: the negative cases asserting a returned `list` is rejected now run with the flag set to `false`; the existing `*_legacy_accept_any_iterable` tests continue to cover the lenient path.
### Was this patch authored or co-authored using generative AI tooling?
No
Closes #58436 from Yicong-Huang/SPARK-58601-default-true.
Authored-by: Yicong Huang <17...@users.noreply.github.com>
Signed-off-by: Yicong-Huang <17...@users.noreply.github.com>
---
python/docs/source/migration_guide/pyspark_upgrade.rst | 4 ----
python/pyspark/sql/tests/arrow/test_arrow_map.py | 11 ++++++++---
python/pyspark/sql/tests/pandas/test_pandas_map.py | 13 +++++++++----
python/pyspark/worker.py | 2 +-
.../main/scala/org/apache/spark/sql/internal/SQLConf.scala | 2 +-
5 files changed, 19 insertions(+), 13 deletions(-)
diff --git a/python/docs/source/migration_guide/pyspark_upgrade.rst b/python/docs/source/migration_guide/pyspark_upgrade.rst
index 35bfe5f23c96..b257ee49b7dd 100644
--- a/python/docs/source/migration_guide/pyspark_upgrade.rst
+++ b/python/docs/source/migration_guide/pyspark_upgrade.rst
@@ -19,10 +19,6 @@
Upgrading PySpark
==================
-Upgrading from PySpark 4.2 to 4.3
----------------------------------
-* In Spark 4.3, a ``mapInPandas`` UDF must return an iterator of ``pandas.DataFrame``\s; returning any other iterable such as a ``list`` now raises ``UDF_RETURN_TYPE``, matching the existing ``mapInArrow`` behavior and the declared ``Iterator[...]`` signature. To restore the previous behavior of accepting any iterable for both ``mapInPandas`` and ``mapInArrow``, set ``spark.sql.execution.pythonUDF.mapInBatch.legacy.acceptAnyIterable.enabled`` to ``true``.
-
Upgrading from PySpark 4.1 to 4.2
---------------------------------
* In Spark 4.2, the minimum supported version for PyArrow has been raised from 15.0.0 to 18.0.0 in PySpark.
diff --git a/python/pyspark/sql/tests/arrow/test_arrow_map.py b/python/pyspark/sql/tests/arrow/test_arrow_map.py
index f2c5644059b6..32525f4fcdd3 100644
--- a/python/pyspark/sql/tests/arrow/test_arrow_map.py
+++ b/python/pyspark/sql/tests/arrow/test_arrow_map.py
@@ -167,9 +167,14 @@ class MapInArrowTestsMixin:
):
(self.spark.range(10, numPartitions=3).mapInArrow(bad_iter_elem, "a int").count())
- with self.assertRaisesRegex(
- PythonException,
- r"iterator of pyarrow\.RecordBatch.*\blist\b",
+ with (
+ self.sql_conf(
+ {"spark.sql.execution.pythonUDF.mapInBatch.legacy.acceptAnyIterable.enabled": False}
+ ),
+ self.assertRaisesRegex(
+ PythonException,
+ r"iterator of pyarrow\.RecordBatch.*\blist\b",
+ ),
):
(self.spark.range(10, numPartitions=3).mapInArrow(list_not_iter, "a int").count())
diff --git a/python/pyspark/sql/tests/pandas/test_pandas_map.py b/python/pyspark/sql/tests/pandas/test_pandas_map.py
index b6cad5b7328e..00ccc06aee94 100644
--- a/python/pyspark/sql/tests/pandas/test_pandas_map.py
+++ b/python/pyspark/sql/tests/pandas/test_pandas_map.py
@@ -232,10 +232,15 @@ class MapInPandasTestsMixin:
):
(self.spark.range(10, numPartitions=3).mapInPandas(bad_iter_elem, "a int").count())
- with self.assertRaisesRegex(
- PythonException,
- "Return type of the user-defined function should be iterator of pandas.DataFrame, "
- "but is list",
+ with (
+ self.sql_conf(
+ {"spark.sql.execution.pythonUDF.mapInBatch.legacy.acceptAnyIterable.enabled": False}
+ ),
+ self.assertRaisesRegex(
+ PythonException,
+ "Return type of the user-defined function should be iterator of pandas.DataFrame, "
+ "but is list",
+ ),
):
(self.spark.range(10, numPartitions=3).mapInPandas(list_not_iter, "a int").count())
diff --git a/python/pyspark/worker.py b/python/pyspark/worker.py
index a37c81893e03..446e90ef843c 100644
--- a/python/pyspark/worker.py
+++ b/python/pyspark/worker.py
@@ -152,7 +152,7 @@ class RunnerConf(Conf):
return (
self.get(
"spark.sql.execution.pythonUDF.mapInBatch.legacy.acceptAnyIterable.enabled",
- "false",
+ "true",
)
== "true"
)
diff --git a/sql/catalyst/src/main/scala/org/apache/spark/sql/internal/SQLConf.scala b/sql/catalyst/src/main/scala/org/apache/spark/sql/internal/SQLConf.scala
index 2e866508c867..f0e6e3b876eb 100644
--- a/sql/catalyst/src/main/scala/org/apache/spark/sql/internal/SQLConf.scala
+++ b/sql/catalyst/src/main/scala/org/apache/spark/sql/internal/SQLConf.scala
@@ -5565,7 +5565,7 @@ object SQLConf {
.version("4.3.0")
.withBindingPolicy(ConfigBindingPolicy.SESSION)
.booleanConf
- .createWithDefault(false)
+ .createWithDefault(true)
val PYTHON_PLANNER_EXEC_MEMORY =
buildConf("spark.sql.planner.pythonExecution.memory")
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@spark.apache.org
For additional commands, e-mail: commits-help@spark.apache.org