Github user JoshRosen commented on a diff in the pull request:
https://github.com/apache/spark/pull/9038#discussion_r41706713
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/execution/aggregate/TungstenAggregationIterator.scala
---
@@ -134,19 +137,74 @@ class TungstenAggregationIterator(
completeAggregateExpressions.map(_.mode).distinct.headOption
}
- // All aggregate functions. TungstenAggregationIterator only handles
expression-based aggregate.
- // If there is any functions that is an ImperativeAggregateFunction, we
throw an
- // IllegalStateException.
- private[this] val allAggregateFunctions: Array[DeclarativeAggregate] = {
- if (!allAggregateExpressions.forall(
- _.aggregateFunction.isInstanceOf[DeclarativeAggregate])) {
- throw new IllegalStateException(
- "Only ExpressionAggregateFunctions should be passed in
TungstenAggregationIterator.")
+ // Initialize all AggregateFunctions by binding references, if necessary,
+ // and setting inputBufferOffset and mutableBufferOffset.
+ private def initializeAllAggregateFunctions(
+ startingInputBufferOffset: Int): Array[AggregateFunction2] = {
+ var mutableBufferOffset = 0
+ var inputBufferOffset: Int = startingInputBufferOffset
+ val functions = new
Array[AggregateFunction2](allAggregateExpressions.length)
+ var i = 0
+ while (i < allAggregateExpressions.length) {
+ val func = allAggregateExpressions(i).aggregateFunction
--- End diff --
I guess that this would also help to avoid potential issues caused by
executor-side mutable state in the aggregate functions, should they wind up
being used multiple times in different parts of a pipeline.
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
