hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Lefty Leverenz (JIRA)" <>
Subject [jira] [Commented] (HIVE-8839) Support "alter table .. add/replace columns cascade"
Date Wed, 19 Nov 2014 08:49:34 GMT


Lefty Leverenz commented on HIVE-8839:

Does this also need to be documented in the Alter Table section of the DDL doc?

* [DDL -- AlterTable/Partition/Column |]
* [DDL -- Add/Replace Columns |]

By the way, CASCADE is only mentioned twice in the DDL doc, in Drop Database and Drop Partitions.
 But Drop Partitions says "For tables that are protected by NO DROP CASCADE, ..." although
NO DROP CASCADE isn't documented anywhere for creating or altering tables.  So it looks like
more doc updates are needed.

* [DDL -- Drop Database |
* [DDL -- Drop Partitions |]

> Support "alter table .. add/replace columns cascade"
> ----------------------------------------------------
>                 Key: HIVE-8839
>                 URL:
>             Project: Hive
>          Issue Type: Improvement
>          Components: SQL
>         Environment: 
>            Reporter: Chaoyu Tang
>            Assignee: Chaoyu Tang
>              Labels: TODOC15
>             Fix For: 0.15.0
>         Attachments: HIVE-8839.1.patch, HIVE-8839.2.patch, HIVE-8839.2.patch, HIVE-8839.patch
> We often run into some issues like HIVE-6131which is due to inconsistent column descriptors
between table and partitions after alter table. HIVE-8441/HIVE-7971 provided the flexibility
to alter table at partition level. But most cases we have need change the table and partitions
at same time. In addition, "alter table" is usually required prior to "alter table partition
.." since querying table partition data is also through table. Instead of do that in two steps,
here we provide a convenient ddl like "alter table ... cascade" to cascade table changes to
partitions as well. The changes are only limited and applicable to add/replace columns and
change column name, datatype, position and comment.

This message was sent by Atlassian JIRA

View raw message