Github user JoshRosen commented on a diff in the pull request:
https://github.com/apache/spark/pull/5725#discussion_r29267684
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/execution/GeneratedAggregate.scala
---
@@ -265,7 +284,49 @@ case class GeneratedAggregate(
val resultProjection = resultProjectionBuilder()
Iterator(resultProjection(buffer))
+ } else if (unsafeEnabled && schemaSupportsUnsafe) {
+ log.info("Using Unsafe-based aggregator")
+ val aggregationMap = new UnsafeFixedWidthAggregationMap(
+ newAggregationBuffer(EmptyRow),
+ aggregationBufferSchema,
+ groupKeySchema,
+ SparkEnv.get.unsafeMemoryManager,
+ 1024 * 16,
--- End diff --
AFAIK we can't use the Scala named argument syntax because
`UnsafeFixedWidthAggregationMap` is implemented in Java, but I can add a line
comment explaining this. Also, I wonder if there's a better choice of default
map size here. If we had statistics / estimates from the optimizer, we could
make a better choice here.
---
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]
