viirya commented on a change in pull request #29467:
URL: https://github.com/apache/spark/pull/29467#discussion_r472312981
##########
File path:
sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/optimizer/ObjectSerializerPruningSuite.scala
##########
@@ -107,4 +108,34 @@ class ObjectSerializerPruningSuite extends PlanTest {
comparePlans(optimized, expected)
}
}
+
+ test("SPARK-32652: Prune nested serializers: RowEncoder") {
+ withSQLConf(SQLConf.SERIALIZER_NESTED_SCHEMA_PRUNING_ENABLED.key ->
"true") {
+ val testRelation = LocalRelation('i.struct(StructType.fromDDL("a int, b
string")), 'j.int)
+ val rowEncoder = RowEncoder(new StructType()
+ .add("i", new StructType().add("a", "int").add("b", "string"))
+ .add("j", "int"))
+ val serializerObject = CatalystSerde.serialize(
+ CatalystSerde.deserialize(testRelation)(rowEncoder))(rowEncoder)
+ val query = serializerObject.select($"i.a")
+ val optimized = Optimize.execute(query.analyze)
+
+ val prunedSerializer = serializerObject.serializer.head.transformDown {
+ case CreateNamedStruct(children) => CreateNamedStruct(children.take(2))
+ }.transformUp {
+ // Aligns null literal in `If` expression to make it resolvable.
+ case i @ If(invoke: Invoke, Literal(null, dt), ser) if
invoke.functionName == "isNullAt" &&
+ !dt.sameType(ser.dataType) =>
+ i.copy(trueValue = Literal(null, ser.dataType))
+ }.asInstanceOf[NamedExpression]
+
+ // `name` in `GetStructField` affects `comparePlans`. Maybe we can ignore
+ // `name` in `GetStructField.equals`?
Review comment:
Sounds good. When we compare `GetStructField` semantically, `name` is
actually ignored.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
