[
https://issues.apache.org/jira/browse/FLINK-40425?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18118067#comment-18118067
]
Zeno Jove commented on FLINK-40425:
-----------------------------------
Hi [~dianfu] , if no one else is working on this, may I take this ticket?
My plan is to follow the API and boundary rules defined in FLIP-591. The
implementation will focus on parameter passing and boundary normalization,
while time attributes and streaming/batch restrictions will continue to be
validated by the planner. timedelta will follow the existing DataFrame API
behavior of truncating values to milliseconds. I will also add the related
tests and documentation.
> Add OVER window aggregation support to DataFrame API
> ----------------------------------------------------
>
> Key: FLINK-40425
> URL: https://issues.apache.org/jira/browse/FLINK-40425
> Project: Flink
> Issue Type: Sub-task
> Components: API / Python
> Reporter: Liu Liu
> Priority: Major
>
> Add OVER-window support to aggregate expressions:
> {{}}
> {code:java}
> def over(
> self,
> order_by,
> *,
> partition_by=None,
> rows=None,
> range=None,
> ) -> Expression {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
