[
https://issues.apache.org/jira/browse/CALCITE-5981?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17762551#comment-17762551
]
Tanner Clary commented on CALCITE-5981:
---------------------------------------
I have done some work in the TimestampDiffConvertlet. It is pretty messy and
whether it's apart of this issue or a separate one, I would love to help clean
it up/help with the review process. I only have experience with BigQuery's
TIMESTAMP_DIFF semantics but that might be relevant, too. Just let me know, and
thanks for filing this issue.
> TIMESTAMPDIFF function returns incorrect result
> ------------------------------------------------
>
> Key: CALCITE-5981
> URL: https://issues.apache.org/jira/browse/CALCITE-5981
> Project: Calcite
> Issue Type: Bug
> Components: core
> Affects Versions: 1.35.0
> Reporter: Mihai Budiu
> Priority: Minor
>
> The following test fails, when added to SqlOperatorTest:
> {code:java}
> f.checkScalar("timestampdiff(month, DATE '2004-02-29', DATE '2005-02-28')",
> "11", "INTEGER NOT NULL");
> {code}
> The result returned by the expression is 12. However, MySQL returns 11.
> The semantics of this function is not described clearly in the documentation,
> but according to several prior issues [1] [2] [3] the intended semantics
> should be the same as in MySQL.
> A corresponding MySQL test: [4]
> The implementation seems to be in
> StandardConvertletTable.TimestampDiffConvertlet.
> [1] https://issues.apache.org/jira/browse/CALCITE-1827,
> [2] https://issues.apache.org/jira/browse/CALCITE-3529,
> [3] https://issues.apache.org/jira/browse/CALCITE-1124.
> [4]
> [https://github.com/mysql/mysql-server/blob/ea1efa9822d81044b726aab20c857d5e1b7e046a/mysql-test/r/func_time.result#L1151]
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
