[
https://issues.apache.org/jira/browse/HIVE-29458?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18077221#comment-18077221
]
Kokila N commented on HIVE-29458:
---------------------------------
Thanks [~difin] for the review and merge.
> Z-order and custom sort expressions in WRITE LOCALLY ORDERED BY cause data
> scatter across reducers due to hash distribution
> ---------------------------------------------------------------------------------------------------------------------------
>
> Key: HIVE-29458
> URL: https://issues.apache.org/jira/browse/HIVE-29458
> Project: Hive
> Issue Type: Bug
> Reporter: Kokila N
> Assignee: Kokila N
> Priority: Major
> Labels: pull-request-available
> Fix For: 4.3.0
>
>
> *Problem*
> When using
> {code:java}
> `WRITE LOCALLY ORDERED BY zorder()`{code}
> or other custom sort expressions on Iceberg tables, the sort expressions are
> added to the ReduceSink partition columns (distribution keys). This causes
> data to be scattered across reducers based on `hash(partition_col, sort_expr)
> % num_reducers`.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
