[IMPALA-11008] Invalid to propagate inferred predicates into the nullable side of an outer join

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: Impala 3.1.0, Impala 3.2.0, Impala 4.0.0, Impala 3.4.0
• Fix Version/s: Impala 4.1.0
• Component/s: Frontend
• Labels: correctness
• Epic Color: ghx-label-3
• Assignee: Xianqing He
• Reporter: Xianqing He
• Created: 08/Nov/21 08:32
• Updated: 04/May/22 10:11
• Resolved: 04/May/22 10:11

Description

Query: explain select * from (
 select case when id is not null and some_nulls is null then true else null end a from (select straight_join t1.id, t2.some_nulls from functional.nullrows t1 left join functional.nullrows t2 on t1.id=t2.id) t)t where a is not null
Max Per-Host Resource Reservation: Memory=1.95MB Threads=3
Per-Host Resource Estimates: Memory=66MB
Codegen disabled by planner
""
PLAN-ROOT SINK
|
02:HASH JOIN [LEFT OUTER JOIN]
|  hash predicates: t1.id = t2.id
|  other predicates: CASE WHEN t1.id IS NOT NULL AND t2.some_nulls IS NULL THEN TRUE ELSE NULL END IS NOT NULL
|  row-size=39B cardinality=26
|
|--01:SCAN HDFS [functional.nullrows t2]
|     partitions=1/1 files=1 size=541B
|     predicates: CASE WHEN t2.id IS NOT NULL AND t2.some_nulls IS NULL THEN TRUE ELSE NULL END IS NOT NULL
|     row-size=26B cardinality=3
|
00:SCAN HDFS [functional.nullrows t1]
   partitions=1/1 files=1 size=541B
   row-size=13B cardinality=26 

It  is incorrect to propagate this inferred predicates into a plan subtree that is on the nullable side of an outer join.

 

Issue Links

• relates to: IMPALA-9162 Incorrect redundant predicate applied to outer join