[
https://issues.apache.org/jira/browse/HUDI-865?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17103910#comment-17103910
]
Balaji Varadarajan commented on HUDI-865:
-----------------------------------------
This is more as a cleanup and to standardize hive schema syncing. The extra
hop is due to recently made change. It also helps to keep hive schema syncing
standardized in future when we support other types like ORC. I don't expect it
to have a performance impact.
> Improve Hive Syncing by directly translating avro schema to Hive types
> ----------------------------------------------------------------------
>
> Key: HUDI-865
> URL: https://issues.apache.org/jira/browse/HUDI-865
> Project: Apache Hudi (incubating)
> Issue Type: Improvement
> Components: Hive Integration
> Reporter: Balaji Varadarajan
> Priority: Major
>
> With the current code in master and proposed improvements with
> [https://github.com/apache/incubator-hudi/pull/1559,|https://github.com/apache/incubator-hudi/pull/1559]
> Hive Sync integration would resort to the following translations for finding
> table schema
> Avro-Schema to Parquet-Schema to Hive Schema transformations
> We need to implement logic to skip the extra hop to parquet schema when
> generating hive schema.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
