[
https://issues.apache.org/jira/browse/HUDI-1550?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17272197#comment-17272197
]
sivabalan narayanan edited comment on HUDI-1550 at 1/26/21, 4:10 PM:
---------------------------------------------------------------------
yeah, if you use the default payload (OverwriteWithLatestAvroPayload), during
merge, Hudi takes in the latest record ignoring the actual ordering value. But
You can define your own implementation for the payload class and achieve this.
Also, in latest release, we have added a new payload implementation called
DefaultHoodieRecordPayload that might come in handy. Ensure you set this config
param "hoodie.payload.ordering.field".
was (Author: shivnarayan):
yeah, if you use the default payload (OverwriteWithLatestAvroPayload), during
merge, Hudi takes in the latest record ignoring the actual ordering value. But
You can define your own implementation for the payload class and achieve this.
Also, in latest release, we have added a new payload implementation called
DefaultHoodieRecordPayload that might come in handy.
> Incorrect query result for MOR table when merge base data with log
> ------------------------------------------------------------------
>
> Key: HUDI-1550
> URL: https://issues.apache.org/jira/browse/HUDI-1550
> Project: Apache Hudi
> Issue Type: Bug
> Components: Spark Integration
> Reporter: pengzhiwei
> Assignee: pengzhiwei
> Priority: Major
> Fix For: 0.8.0
>
>
> Table A is (id: int, value: string, ts: long), where "id" is the record key,
> "ts" is the precombine key. Update table A which the follow data:
> {code:java}
> (1, '10', 12)
> (1,'11', 10){code}
>
> The result of "select * from A where id = 1" should be *(1, '10', 12)*.
> However hoodie currently return *(1, '11', 10)* which is not the right answer.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
