[
https://issues.apache.org/jira/browse/HIVE-28658?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17905365#comment-17905365
]
Butao Zhang commented on HIVE-28658:
------------------------------------
[~dkuzmenko] Implementing the multi-catalog feat like Trino is definitely a
good direction!
Let me make an aside. I know that Apache Doris & Trino & Apache Spark are all
have the multi-catalog feat, which can make the multi datasource federated
query more flexible. {color:#de350b}At present, in HS2, query on other
datasources such as jdbc(mysql) must go through HMS{color}. We can think about
decoupling HS2 and HMS, and then HS2 can be like Trino to query different
datasource without going through HMS. Certainly, this can let HS2 to query rest
iceberg table more easier. Just like this you post:
{code:java}
iceberg.catalog.type=rest
iceberg.rest-catalog.uri=https://iceberg-with-rest:8181/{code}
> Iceberg REST Catalog Support
> -----------------------------
>
> Key: HIVE-28658
> URL: https://issues.apache.org/jira/browse/HIVE-28658
> Project: Hive
> Issue Type: Improvement
> Security Level: Public(Viewable by anyone)
> Components: Iceberg integration
> Reporter: Denys Kuzmenko
> Assignee: Zoltán Rátkai
> Priority: Major
>
> Allow Hive to communicate directly with an Iceberg catalog service using the
> Iceberg REST Catalog protocol. This protocol is the standard and recommended
> way to interact with Iceberg metadata and tables.
> interface MetadataProvider
> * HiveCatalog
> * RestCatalog
> Ref:
> https://www.tabular.io/apache-iceberg-cookbook/getting-started-trino-configuration/
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
