This is an automated email from the ASF dual-hosted git repository.
wenchen pushed a commit to branch branch-3.5
in repository https://gitbox.apache.org/repos/asf/spark.git
The following commit(s) were added to refs/heads/branch-3.5 by this push:
new e3317a7cc66c [SPARK-47385] Fix tuple encoders with Option inputs
e3317a7cc66c is described below
commit e3317a7cc66cc9ce893f54ebf19a038d32026c4a
Author: Chenhao Li <chenhao...@databricks.com>
AuthorDate: Thu Mar 14 14:27:36 2024 +0800
[SPARK-47385] Fix tuple encoders with Option inputs
## What changes were proposed in this pull request?
https://github.com/apache/spark/pull/40755 adds a null check on the input
of the child deserializer in the tuple encoder. It breaks the deserializer for
the `Option` type, because null should be deserialized into `None` rather than
null. This PR adds a boolean parameter to `ExpressionEncoder.tuple` so that
only the user that https://github.com/apache/spark/pull/40755 intended to fix
has this null check.
## How was this patch tested?
Unit test.
Closes #45508 from chenhao-db/SPARK-47385.
Authored-by: Chenhao Li <chenhao...@databricks.com>
Signed-off-by: Wenchen Fan <wenc...@databricks.com>
(cherry picked from commit 9986462811f160eacd766da8a4e14a9cbb4b8710)
Signed-off-by: Wenchen Fan <wenc...@databricks.com>
---
.../spark/sql/catalyst/encoders/ExpressionEncoder.scala | 10 ++++++++--
sql/core/src/main/scala/org/apache/spark/sql/Dataset.scala | 4 +++-
.../src/test/scala/org/apache/spark/sql/DatasetSuite.scala | 12 ++++++++++++
3 files changed, 23 insertions(+), 3 deletions(-)
diff --git
a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/encoders/ExpressionEncoder.scala
b/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/encoders/ExpressionEncoder.scala
index ff72b5a0d965..6dc89bb4d4b1 100644
---
a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/encoders/ExpressionEncoder.scala
+++
b/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/encoders/ExpressionEncoder.scala
@@ -73,8 +73,14 @@ object ExpressionEncoder {
* Given a set of N encoders, constructs a new encoder that produce objects
as items in an
* N-tuple. Note that these encoders should be unresolved so that
information about
* name/positional binding is preserved.
+ * When `useNullSafeDeserializer` is true, the deserialization result for a
child will be null if
+ * the input is null. It is false by default as most deserializers handle
null input properly and
+ * don't require an extra null check. Some of them are null-tolerant, such
as the deserializer for
+ * `Option[T]`, and we must not set it to true in this case.
*/
- def tuple(encoders: Seq[ExpressionEncoder[_]]): ExpressionEncoder[_] = {
+ def tuple(
+ encoders: Seq[ExpressionEncoder[_]],
+ useNullSafeDeserializer: Boolean = false): ExpressionEncoder[_] = {
if (encoders.length > 22) {
throw QueryExecutionErrors.elementsOfTupleExceedLimitError()
}
@@ -119,7 +125,7 @@ object ExpressionEncoder {
case GetColumnByOrdinal(0, _) => input
}
- if (enc.objSerializer.nullable) {
+ if (useNullSafeDeserializer && enc.objSerializer.nullable) {
nullSafe(input, childDeserializer)
} else {
childDeserializer
diff --git a/sql/core/src/main/scala/org/apache/spark/sql/Dataset.scala
b/sql/core/src/main/scala/org/apache/spark/sql/Dataset.scala
index c063af9381ff..9ef48c2ec42e 100644
--- a/sql/core/src/main/scala/org/apache/spark/sql/Dataset.scala
+++ b/sql/core/src/main/scala/org/apache/spark/sql/Dataset.scala
@@ -1229,7 +1229,9 @@ class Dataset[T] private[sql](
JoinHint.NONE)).analyzed.asInstanceOf[Join]
implicit val tuple2Encoder: Encoder[(T, U)] =
- ExpressionEncoder.tuple(this.exprEnc, other.exprEnc)
+ ExpressionEncoder
+ .tuple(Seq(this.exprEnc, other.exprEnc), useNullSafeDeserializer =
true)
+ .asInstanceOf[Encoder[(T, U)]]
withTypedPlan(JoinWith.typedJoinWith(
joined,
diff --git a/sql/core/src/test/scala/org/apache/spark/sql/DatasetSuite.scala
b/sql/core/src/test/scala/org/apache/spark/sql/DatasetSuite.scala
index c2fe31520acf..f32b32ffc5a5 100644
--- a/sql/core/src/test/scala/org/apache/spark/sql/DatasetSuite.scala
+++ b/sql/core/src/test/scala/org/apache/spark/sql/DatasetSuite.scala
@@ -2504,6 +2504,18 @@ class DatasetSuite extends QueryTest
assert(result == expected)
}
+ test("SPARK-47385: Tuple encoder with Option inputs") {
+ implicit val enc: Encoder[(SingleData, Option[SingleData])] =
+ Encoders.tuple(Encoders.product[SingleData],
Encoders.product[Option[SingleData]])
+
+ val input = Seq(
+ (SingleData(1), Some(SingleData(1))),
+ (SingleData(2), None)
+ )
+ val ds = spark.createDataFrame(input).as[(SingleData, Option[SingleData])]
+ checkDataset(ds, input: _*)
+ }
+
test("SPARK-43124: Show does not trigger job execution on CommandResults") {
withSQLConf(SQLConf.OPTIMIZER_EXCLUDED_RULES.key -> "") {
withTable("t1") {
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@spark.apache.org
For additional commands, e-mail: commits-h...@spark.apache.org
