Prashant Singh created SPARK-39678:
--------------------------------------
Summary: Improve stats estimation for v2 tables
Key: SPARK-39678
URL: https://issues.apache.org/jira/browse/SPARK-39678
Project: Spark
Issue Type: Improvement
Components: Optimizer
Affects Versions: 3.3.0
Reporter: Prashant Singh
In case of v2 tables, connectors can bubble up both [sizeInBytes and rowCount
|https://github.com/apache/spark/blob/master/sql/catalyst/src/main/java/org/apache/spark/sql/connector/read/Statistics.java].
Presently, SizeInBytesOnlyStatsPlanVisitor, ommits propagating / estimating
rowCount stats, some places like :
*
[CodePointer1|https://github.com/apache/spark/blob/master/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/statsEstimation/SizeInBytesOnlyStatsPlanVisitor.scala#L54-L58]
* [CodePointer2
|https://github.com/apache/spark/blob/master/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/statsEstimation/SizeInBytesOnlyStatsPlanVisitor.scala#L46-L47]
For the
[non-cbo|https://github.com/apache/spark/blob/master/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/statsEstimation/LogicalPlanStats.scala#L34-L39]
flow, as per my understanding, this can improve the stats estimation, since
rowcount is indirectly used in places to estimate the size as well.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
