[SPARK-37828] Push down filters through RebalancePartitions

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.3.0
• Fix Version/s: 3.3.0
• Component/s: SQL
• Assignee: Yuming Wang
• Reporter: Yuming Wang
• Created: 06/Jan/22 10:11
• Updated: 10/Jan/22 01:00
• Resolved: 10/Jan/22 01:00

Description

How to reproduce this issue:

spark.sql("SELECT * FROM (SELECT /*+ REBALANCE */  * from range(10)) t11 WHERE id = 3").explain(true)

Output:

== Optimized Logical Plan ==
Filter (id#0L = 3)
+- RebalancePartitions
   +- Range (0, 10, step=1, splits=None)

Expected:

== Optimized Logical Plan ==
RebalancePartitions
+- Filter (id#0L = 3)
   +- Range (0, 10, step=1, splits=None)

Issue Links

• links to: [Github] Pull Request #35118 (wangyum)
• links to: [Github] Pull Request #35118 (wangyum)