[ https://issues.apache.org/jira/browse/SQOOP-3323?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16520178#comment-16520178
]
Daniel Voros commented on SQOOP-3323:
-------------------------------------
Attached review request.
> Use hive executable in (non-JDBC) Hive imports
> ----------------------------------------------
>
> Key: SQOOP-3323
> URL: https://issues.apache.org/jira/browse/SQOOP-3323
> Project: Sqoop
> Issue Type: Improvement
> Components: hive-integration
> Affects Versions: 3.0.0
> Reporter: Daniel Voros
> Assignee: Daniel Voros
> Priority: Major
> Fix For: 3.0.0
>
>
> When doing Hive imports the old way (not via JDBC that was introduced in SQOOP-3309)
we're trying to use the {{CliDriver}} class from Hive and fall back to the {{hive}} executable
(a.k.a. [Hive Cli|https://cwiki.apache.org/confluence/display/Hive/LanguageManual+Cli]) if
that class is not found.
> Since {{CliDriver}} and the {{hive}} executable that's relying on it are [deprecated|https://cwiki.apache.org/confluence/display/Hive/LanguageManual+Cli]
(see also HIVE-10511), we should switch to using {{beeline}} to talk to Hive. With recent
additions (e.g. HIVE-18963) this should be easier than before.
> As a first step we could switch to using {{hive}} executable. With HIVE-19728 it will
be possible (in Hive 3.1) to configure hive to actually run beeline when using the {{hive}}
executable. This way we could leave it to the user to decide whether to use the deprecated
cli or use beeline instead.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
|