flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Rui Li (Jira)" <j...@apache.org>
Subject [jira] [Commented] (FLINK-22272) Some scenes can't drop table by hive catalog
Date Mon, 31 May 2021 03:16:00 GMT

    [ https://issues.apache.org/jira/browse/FLINK-22272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17354192#comment-17354192

Rui Li commented on FLINK-22272:

Hi [~wangzeyu], could you explain what's the benefit to move the feature from CatalogManager
to Catalog? As a SQL engine, Flink should provide consistent semantics for DROP TABLE/VIEW
no matter what Catalog is being used. So moving it to Catalog means each implementation needs
to repeat this logic, right? Besides, it requires public API change which has bigger impact
on existing users and should be discussed on the mailing list.

As to #2, as I said it has noting to do with DROP TABLE/VIEW. We should fix it anyway.

> Some scenes can't drop table by hive catalog
> --------------------------------------------
>                 Key: FLINK-22272
>                 URL: https://issues.apache.org/jira/browse/FLINK-22272
>             Project: Flink
>          Issue Type: Bug
>          Components: Connectors / Hive
>    Affects Versions: 1.12.1
>            Reporter: wangzeyu
>            Priority: Major
>              Labels: pull-request-available, stale-major
>             Fix For: 1.12.1
> There are some scenes we can't drop table by hive catalog because of we execute error
create table sql like "caeate table tableName". And then when we execute drop table sql "drop
table tableName" will throw error "org.apache.flink.table.catalog.exceptions.CatalogException:
Failed to get table schema from properties for generic table".
> So, i think maybe we should remove the check when we drop table by third-party catalog
and just drop table depend on the third-party catalog.

This message was sent by Atlassian Jira

View raw message