You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@flink.apache.org by "Jane Chan (Jira)" <ji...@apache.org> on 2024/01/04 02:49:00 UTC

[jira] [Commented] (FLINK-33941) Use field reference index to compute window aggregate time attribute column

    [ https://issues.apache.org/jira/browse/FLINK-33941?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17802374#comment-17802374 ] 

Jane Chan commented on FLINK-33941:
-----------------------------------

Fixed in master: b957480112c00d9d777247fc48b602e9908652a2

> Use field reference index to compute window aggregate time attribute column
> ---------------------------------------------------------------------------
>
>                 Key: FLINK-33941
>                 URL: https://issues.apache.org/jira/browse/FLINK-33941
>             Project: Flink
>          Issue Type: Technical Debt
>          Components: Table SQL / Planner
>            Reporter: xuyang
>            Assignee: xuyang
>            Priority: Minor
>              Labels: pull-request-available
>
> In some exec nodes like StreamExecGroupWindowAggregate and some rules like BatchPhysicalWindowAggregateRule, planner uses "AggregateUtil#timeFieldIndex" to access the actual time field index, instead of using the time field index in LogicalWindow#timeAttribute directly. However, it would be more formal to use the field index instead of the column field.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)