[
https://issues.apache.org/jira/browse/NIFI-4963?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16408610#comment-16408610
]
Matt Burgess commented on NIFI-4963:
------------------------------------
I think we should replace ConvertAvroToORC in the Hive 3 bundle with a PutORC
processor, it would basically be a mashup of ConvertAvroToORC and PutParquet,
so it would be record-aware, would write directly to HDFS, but can still
generate DDL in attribute(s) for later commands for PutHiveQL (or PutHive3QL as
it were). We can still use Hive's OrcFile and Writer interfaces (rather than
pure Apache ORC), that should make the upgrade easier, and Hive 3's ORC
interfaces are thin veneers over the Apache ORC stuff anyway.
I don't like to necessarily add "3" to the processor names, but I think we need
to call them something different so there is no user confusion on which
processor to pick. I'm relegated to PutHive3QL, SelectHive3QL, and
Hive3ConnectionPool unless there are other suggestions?
> Add support for Hive 3.0 processors
> -----------------------------------
>
> Key: NIFI-4963
> URL: https://issues.apache.org/jira/browse/NIFI-4963
> Project: Apache NiFi
> Issue Type: New Feature
> Components: Extensions
> Reporter: Matt Burgess
> Assignee: Matt Burgess
> Priority: Major
>
> Apache Hive is working on Hive 3.0, this Jira is to add a bundle of
> components (much like the current Hive bundle) that supports Hive 3.0 (and
> Apache ORC if necessary).
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
