Aihua Xu updated HIVE-14926:
    Status: Patch Available  (was: Open)

Patch-1: switch the query execution from beeline to use JDBC directly. This 
doesn't seem to be necessary since internally it goes through JDBC as well but 
limits the supported functionalities.

Switch to call JDBC connection directly with which we can use batch and commit 
or rollback the change. Also, now dryRun option is doing the real dryRun which 
will execute the scripts and give you the execution info without the final 

> Keep Schema in consistent state where schemaTool fails or succeeds.  
> ---------------------------------------------------------------------
>                 Key: HIVE-14926
>                 URL: https://issues.apache.org/jira/browse/HIVE-14926
>             Project: Hive
>          Issue Type: Improvement
>          Components: Database/Schema
>            Reporter: Aihua Xu
>            Assignee: Aihua Xu
>         Attachments: HIVE-14926.1.patch
> SchemaTool uses autocommit right now when executing the upgrade or init 
> scripts. Seems we should use database transaction to commit or roll back to 
> keep schema consistent.  

This message was sent by Atlassian JIRA

Reply via email to