[
https://issues.apache.org/jira/browse/CALCITE-5118?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17528852#comment-17528852
]
Julian Hyde commented on CALCITE-5118:
--------------------------------------
I haven’t looked closely at this but I don’t think that the validator should be
circumvented. In all cases there should be a validation error with a position.
Maybe that means that we trust people to validate before calling rewriteCall.
> SqlDatePartFunction#rewriteCall should check operands length
> ------------------------------------------------------------
>
> Key: CALCITE-5118
> URL: https://issues.apache.org/jira/browse/CALCITE-5118
> Project: Calcite
> Issue Type: Improvement
> Components: core
> Affects Versions: 1.30.0
> Reporter: Benchao Li
> Assignee: Benchao Li
> Priority: Major
> Labels: pull-request-available
> Fix For: 1.31.0
>
> Time Spent: 40m
> Remaining Estimate: 0h
>
> {{SqlDatePartFunction#rewriteCall}} does not check operands length for now,
> and it will be rewrite to {{EXTRACT}} unconditionally before validation.
> E.g.
> {code:sql}
> select week(); -- will throw ArrayIndexOutOfBoundsException
> select week(date '2022-04-27', 1); -- will just ignore second param
> {code}
> However, these cases should throw validation exceptions, because {{week}}
> actually only accept 1 operand.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
