You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "ASF GitHub Bot (Jira)" <ji...@apache.org> on 2026/10/07 20:14:00 UTC
[jira] [Updated] (FLINK-40949) WindowedUnsliceAssigner reports a NULL window bound as "RowTime field should not be null."
[ https://issues.apache.org/jira/browse/FLINK-40949?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
ASF GitHub Bot updated FLINK-40949:
-----------------------------------
Labels: pull-request-available (was: )
> WindowedUnsliceAssigner reports a NULL window bound as "RowTime field should not be null."
> ------------------------------------------------------------------------------------------
>
> Key: FLINK-40949
> URL: https://issues.apache.org/jira/browse/FLINK-40949
> Project: Flink
> Issue Type: Bug
> Components: Table SQL / Planner
> Affects Versions: 1.19.3, 2.3.0, 2.2.1, 1.20.5, 2.4.0
> Reporter: Martijn Visser
> Priority: Major
> Labels: pull-request-available
>
> {{UnsliceAssigners.WindowedUnsliceAssigner#createWindow}} throws {{RuntimeException("RowTime field should not be null.")}} when window_start or window_end of an incoming row is NULL (added with FLINK-34100). The text is misleading: this assigner only runs on rows that an upstream window operator produced, where both columns are TIMESTAMP(3) NOT NULL, so a NULL here means an operator or planner defect and not a NULL in the user's rowtime column. {{WatermarkAssignerOperator}} and {{SliceAssigners.AbstractSliceAssigner}} use the same text for a NULL rowtime attribute, which is a data problem the user can fix, and anything that classifies failures by message cannot tell the two apart. {{SliceAssigners.WindowedSliceAssigner}} has no null check at all for the same situation.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)