[ 
https://issues.apache.org/jira/browse/PHOENIX-6098?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17183551#comment-17183551
 ] 

Josh Elser commented on PHOENIX-6098:
-------------------------------------

{quote}It's very hard to add tests to phoenix-hive module of 
phoenix-connectors, so I didn't add any tests
{quote}
Do you have an example SQL query which would could be validated by hand? I can 
see how the code is broken given the context you've shared, but i'm not sure 
how I would write such a query to demonstrate that Phoenix-Hive does the 
correct thing after your change :)

> IndexPredicateAnalyzer wrongly handles pushdown predicates and residual 
> predicates
> ----------------------------------------------------------------------------------
>
>                 Key: PHOENIX-6098
>                 URL: https://issues.apache.org/jira/browse/PHOENIX-6098
>             Project: Phoenix
>          Issue Type: Bug
>          Components: hive-connector
>            Reporter: Toshihiro Suzuki
>            Assignee: Toshihiro Suzuki
>            Priority: Major
>
> Currently, the following code of IndexPredicateAnalyzer is assuming that 
> GenericUDFOPAnd always has 2 children nodes. I think this is wrong and it 
> leads wrong results:
> https://github.com/apache/phoenix-connectors/blob/5bd23ae2a0f70c3b3edf92a53780dafa643faf26/phoenix-hive/src/main/java/org/apache/phoenix/hive/ql/index/IndexPredicateAnalyzer.java#L346-L359



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to