[
https://issues.apache.org/jira/browse/HIVE-14226?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15375130#comment-15375130
]
Barna Zsombor Klara commented on HIVE-14226:
--------------------------------------------
Hi [~jcamachorodriguez],
I would like to provide a meaningful unit test for this case, but I am a bit
lost in the CalcitePlanner.
My first approach was to generate a simple syntax tree containing a
TOK_FUNCTIONDI, then set it on the QueryBlock of a CalcitePlanner, and call the
CalcitePlanner#getOptimizedAST, but it seems the QueryBlock needs a lot more
setup than just an ASTNode to work properly.
Since you are handling CBO jiras, could you maybe help me with a better
approach?
Alternatively I can provide a unit test of an ASTNode showing that
TOK_FUNCTIONDI and other tokens are indeed turned lowercase in the string tree.
> Invalid check on an ASTNode#toStringTree in CalcitePlanner
> ----------------------------------------------------------
>
> Key: HIVE-14226
> URL: https://issues.apache.org/jira/browse/HIVE-14226
> Project: Hive
> Issue Type: Bug
> Reporter: Barna Zsombor Klara
> Priority: Minor
>
> The patch for HIVE-12590 has modified the string tree representation of an
> ASTNode, so that everything except string literals are now lowercase.
> However CalcitePlanner#genSelectLogicalPlan is still checking the occurrence
> of an uppercase TOK_FUNCTIONDI constant in the string tree.
> (Line 2997 in CalcitePlanner on master)
> I think this check will never evaluate to true, so we should either be using
> a lowercase String, or remove the check if it is unnecessary.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
