[
https://issues.apache.org/jira/browse/HUDI-9076?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17930249#comment-17930249
]
Geser Dugarov edited comment on HUDI-9076 at 2/26/25 4:45 AM:
--------------------------------------------------------------
But there is corresponding task to enable it for Flink write by default:
[HUDI-8814] Enable col stats and partition stats by default on the writer with
flink
https://issues.apache.org/jira/browse/HUDI-8814
Need to figure out which behavior is expected.
was (Author: JIRAUSER301110):
But there is corresponding task to enable it for Flink write by default:
[HUDI-8814] Enable col stats and partition stats by default on the writer with
flink
https://issues.apache.org/jira/browse/HUDI-8814
Need to figure out what behavior is expected.
> Enabling cols stats by default with writer for Spark enabled it for Flink
> -------------------------------------------------------------------------
>
> Key: HUDI-9076
> URL: https://issues.apache.org/jira/browse/HUDI-9076
> Project: Apache Hudi
> Issue Type: Bug
> Reporter: Geser Dugarov
> Assignee: Geser Dugarov
> Priority: Major
>
> [HUDI-8766] Enabling cols stats by default with writer (#12596)
> - Enabling cols stats by default on writer in SPARK engine
> - Added support for timestamp, Date, LocalDate, Decimal.
> was expected to turn on column stats for Spark engine.
> But I found performance decrease on this commit for Flink write due to
> enabled column stats.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
