[
https://issues.apache.org/jira/browse/HIVE-12223?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jesus Camacho Rodriguez updated HIVE-12223:
-------------------------------------------
Description:
Consider the following query:
{noformat}
SELECT key, value, GROUPING__ID, count(*)
FROM T1
GROUP BY key, value
GROUPING SETS ((), (key))
HAVING GROUPING__ID = 1
{noformat}
This query will not return results.
The reason is that a "constant" placeholder is introduced by SemanticAnalyzer
for the GROUPING\__ID column. At execution time, this placeholder is replaced
by the actual value of the GROUPING\__ID. As the column is a constant, the Hive
optimizer (combination of PPD and constant folding) will evaluate statically
whether the condition is met or not, leading to incorrect results.
We should be able to recognize the placeholder and avoid PPD pushing the filter
predicate to the GHive optimizer evaluates it statically.
was:
Consider the following query:
{noformat}
SELECT key, value, GROUPING__ID, count(*)
FROM T1
GROUP BY key, value
GROUPING SETS ((), (key))
HAVING GROUPING__ID = 1
{noformat}
This query will not return results.
The reason is that a "constant" placeholder is introduced by SemanticAnalyzer
for the GROUPING\__ID column. At execution time, this placeholder is replaced
by the actual value of the GROUPING\__ID. As the column is a constant, the Hive
optimizer (combination of PPD and constant folding) will evaluate statically
whether the condition is met or not, leading to incorrect results.
We should be able to recognize the placeholder and avoid that PPD pushes the
filter predicate to the GHive optimizer evaluates it statically.
> Filter on Grouping__ID does not work properly
> ---------------------------------------------
>
> Key: HIVE-12223
> URL: https://issues.apache.org/jira/browse/HIVE-12223
> Project: Hive
> Issue Type: Bug
> Components: Logical Optimizer
> Affects Versions: 1.3.0, 2.0.0
> Reporter: Jesus Camacho Rodriguez
> Assignee: Jesus Camacho Rodriguez
> Fix For: 1.3.0, 2.0.0
>
> Attachments: HIVE-12223.01.patch, HIVE-12223.patch
>
>
> Consider the following query:
> {noformat}
> SELECT key, value, GROUPING__ID, count(*)
> FROM T1
> GROUP BY key, value
> GROUPING SETS ((), (key))
> HAVING GROUPING__ID = 1
> {noformat}
> This query will not return results.
> The reason is that a "constant" placeholder is introduced by SemanticAnalyzer
> for the GROUPING\__ID column. At execution time, this placeholder is replaced
> by the actual value of the GROUPING\__ID. As the column is a constant, the
> Hive optimizer (combination of PPD and constant folding) will evaluate
> statically whether the condition is met or not, leading to incorrect results.
> We should be able to recognize the placeholder and avoid PPD pushing the
> filter predicate to the GHive optimizer evaluates it statically.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
