[
https://issues.apache.org/jira/browse/SPARK-58823?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Uroš Bojanić updated SPARK-58823:
---------------------------------
Component/s: SQL
Tests
(was: Spark Core)
> Add nanos-timestamp tests for scalar subquery / EXISTS / NOT IN (SELECT)
> ------------------------------------------------------------------------
>
> Key: SPARK-58823
> URL: https://issues.apache.org/jira/browse/SPARK-58823
> Project: Spark
> Issue Type: Task
> Components: SQL, Tests
> Affects Versions: 4.3.0
> Reporter: Stevo Mitric
> Assignee: Stevo Mitric
> Priority: Major
> Labels: pull-request-available
>
> Scalar subquery, EXISTS/NOT EXISTS, and NOT IN (SELECT) over
> nanosecond-precision timestamps have no explicit test coverage — only IN
> (SELECT) is tested. They work generically (shared nanos equality and type
> widening), but nothing locks it, so a future change could truncate
> sub-microsecond values to micros undetected. Add end-to-end coverage
> (test-only, no production change): SQL golden cases in
> timestamp-ntz-nanos.sql plus a DataFrame test in
> TimestampNanosWideningSuiteBase (ANSI on/off), covering scalar subquery
> (value/typeof/NULL, and in a WHERE compare), correlated EXISTS/NOT EXISTS,
> and NOT IN including mixed-precision (p=7 vs p=9) and the NULL-in-set
> three-valued-logic case.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
