[
https://issues.apache.org/jira/browse/HIVE-617?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Edward Capriolo updated HIVE-617:
---------------------------------
Attachment: hive-617.3.diff
This patch specifies the jarfile and class name as command line arguments like
hadoop does. With this change the ordering of jarfile, classname and -hiveconf
is now significant. Launching a jar from hadoop has similar constraints so this
should not be an issue.
> Script to start classes with hadoop and hive environment
> --------------------------------------------------------
>
> Key: HIVE-617
> URL: https://issues.apache.org/jira/browse/HIVE-617
> Project: Hadoop Hive
> Issue Type: New Feature
> Components: Clients
> Reporter: Edward Capriolo
> Assignee: Edward Capriolo
> Priority: Minor
> Attachments: hive-617.2.patch, hive-617.3.diff, hive-617.patch,
> TestHive.java
>
>
> At times it may be required to write a process that uses both the Hadoop and
> Hive environment and API. For example, someone may write an application that
> uses the HIVE api directly. This patch will add a more generic --jar
> extension that can start any class with the proper environment.
> RUNJAR=/opt/hive/lib/hive_hwi.jar RUNCLASS=test.TestHive hive --service jar
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
