Github user rxin commented on a diff in the pull request:
https://github.com/apache/spark/pull/9788#discussion_r45167423
--- Diff: core/src/main/scala/org/apache/spark/util/SizeEstimator.scala ---
@@ -32,6 +32,16 @@ import org.apache.spark.annotation.DeveloperApi
import org.apache.spark.util.collection.OpenHashSet
/**
+ * A trait that allows a class to give [[SizeEstimator]] more accurate
size estimation.
+ * When a class extends it, [[SizeEstimator]] will query the
`estimatedSize` first.
+ * If `estimatedSize` does not return [[None]], [[SizeEstimator]] will use
the returned size
+ * as the size of the object. Otherwise, [[SizeEstimator]] will do the
estimation work.
+ */
+private[spark] trait SizeEstimation {
+ def estimatedSize: Option[Long]
--- End diff --
The high level approach of not relying on reflection and object walking is
a good one -- actually with dataset and dataframes, we don't really need size
estimation. I also think relying on thread locals and SparkEnv is much less
ideal than explicit dependencies.
Either way, this pull request is ok to merge in its current shape, given
it's fairly critical. We can do more changes later.
---
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]
