Quanlong Huang created IMPALA-15757:
---------------------------------------
Summary: TestHBO.test_join_over_analytic_cardinality fails in
calcite-planner builds
Key: IMPALA-15757
URL: https://issues.apache.org/jira/browse/IMPALA-15757
Project: IMPALA
Issue Type: Bug
Reporter: Quanlong Huang
Assignee: Quanlong Huang
TestHBO.test_join_over_analytic_cardinality fails in calcite-planner builds due
to calcite-planner generates additional SELECT nodes.
{noformat}
Expected (Original Planner): Actual
(Calcite Planner):
PLAN-ROOT SINK PLAN-ROOT
SINK
| |
06:HASH JOIN [INNER JOIN, BROADCAST] 08:HASH JOIN
[INNER JOIN, BROADCAST]
| hash predicates: int_col = int_col, rank() = rank() | hash
predicates: int_col = int_col, rank() = rank()
| row-size=58B cardinality=0 (from HBO) |
row-size=58B cardinality=62
| |
|--09:EXCHANGE [UNPARTITIONED]
|--11:EXCHANGE [UNPARTITIONED]
| | | |
| 05:ANALYTIC | 07:SELECT
| | functions: rank() | |
predicates: int_col IS NOT NULL
| | order by: id DESC | |
row-size=29B cardinality=1
| | window: RANGE BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW | |
| | row-size=29B cardinality=2 |
06:ANALYTIC
| | | |
functions: rank()
| 08:MERGING-EXCHANGE [UNPARTITIONED] | | order
by: id DESC
| | order by: id DESC | |
window: RANGE BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW
| | | |
row-size=29B cardinality=2
| 04:SORT | |
| | order by: id DESC |
10:MERGING-EXCHANGE [UNPARTITIONED]
| | row-size=21B cardinality=2 | | order
by: id DESC NULLS FIRST
| | | |
| 03:SCAN HDFS [functional.alltypestiny] | 05:SORT
| partition predicates: `month` = 1 | | order
by: id DESC NULLS FIRST
| HDFS partitions=1/4 files=1 size=115B | |
row-size=21B cardinality=2
| predicates: string_col != 'test_join_over_analytic' | |
| row-size=21B cardinality=2 (from HBO) | 04:SCAN
HDFS [functional.alltypestiny]
| |
partition predicates: functional.alltypestiny.month = 1
02:ANALYTIC | HDFS
partitions=1/4 files=1 size=115B
| functions: rank() |
predicates: functional.alltypestiny.string_col != 'test_join_over_analytic'
| order by: id ASC |
row-size=21B cardinality=2 (from HBO)
| window: RANGE BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW |
| row-size=29B cardinality=620 03:SELECT
| |
predicates: int_col IS NOT NULL
07:MERGING-EXCHANGE [UNPARTITIONED] |
row-size=29B cardinality=62
| order by: id ASC |
| 02:ANALYTIC
01:SORT |
functions: rank()
| order by: id ASC | order by:
id ASC
| row-size=21B cardinality=620 | window:
RANGE BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW
| |
row-size=29B cardinality=620
00:SCAN HDFS [functional.alltypes] |
partition predicates: `month` = 1
09:MERGING-EXCHANGE [UNPARTITIONED]
HDFS partitions=2/24 files=2 size=40.32KB | order by:
id ASC NULLS LAST
predicates: string_col != 'test_join_over_analytic' |
row-size=21B cardinality=620 (from HBO) 01:SORT
~ | order by:
id ASC NULLS LAST
~ |
row-size=21B cardinality=620
~ |
~ 00:SCAN HDFS
[functional.alltypes]
~ partition
predicates: functional.alltypes.month = 1
~ HDFS
partitions=2/24 files=2 size=40.32KB
~
predicates: functional.alltypes.string_col != 'test_join_over_analytic'
~
row-size=21B cardinality=620 (from HBO){noformat}
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
