Github user cloud-fan commented on a diff in the pull request:
https://github.com/apache/spark/pull/16371#discussion_r94114431
--- Diff:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/aggregate/collect.scala
---
@@ -44,39 +44,78 @@ abstract class Collect extends ImperativeAggregate {
override def dataType: DataType = ArrayType(child.dataType)
- override def supportsPartial: Boolean = false
-
- override def aggBufferAttributes: Seq[AttributeReference] = Nil
-
- override def aggBufferSchema: StructType =
StructType.fromAttributes(aggBufferAttributes)
-
- override def inputAggBufferAttributes: Seq[AttributeReference] = Nil
-
// Both `CollectList` and `CollectSet` are non-deterministic since their
results depend on the
// actual order of input rows.
override def deterministic: Boolean = false
- protected[this] val buffer: Growable[Any] with Iterable[Any]
-
- override def initialize(b: InternalRow): Unit = {
- buffer.clear()
+ private def generateOutput(results: Iterable[Any]): Any = {
+ if (results.isEmpty) {
+ null
+ } else {
+ new GenericArrayData(results.toArray)
+ }
}
- override def update(b: InternalRow, input: InternalRow): Unit = {
+ override def update(buffer: T, input: InternalRow): T = {
+ val value = child.eval(input).asInstanceOf[Any]
+
// Do not allow null values. We follow the semantics of Hive's
collect_list/collect_set here.
// See:
org.apache.hadoop.hive.ql.udf.generic.GenericUDAFMkCollectionEvaluator
- val value = child.eval(input)
if (value != null) {
buffer += value
}
+ buffer
+ }
+
+ override def merge(buffer: T, other: T): T = {
+ buffer ++= other
}
- override def merge(buffer: InternalRow, input: InternalRow): Unit = {
- sys.error("Collect cannot be used in partial aggregations.")
+ override def eval(buffer: T): Any = {
+ generateOutput(buffer)
}
- override def eval(input: InternalRow): Any = {
- new GenericArrayData(buffer.toArray)
+ protected def _serialize(obj: Iterable[Any]): Array[Byte] = {
+ val buffer = new Array[Byte](4 << 10) // 4K
--- End diff --
shall we create `new GenericArrayData(buffer.toArray)` and do unsafe
projection?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
