[
https://issues.apache.org/jira/browse/HIVE-7446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14088006#comment-14088006
]
Szehon Ho commented on HIVE-7446:
---------------------------------
I think the value of 'alter table' scenario is not to override the first set of
data, as in your last statement, but instead the user wants to use both set of
data.
Otherwise this example (and the one in your test case) can also be achieved by
just dropping doctor table and recreating.
{noformat}
create table doctors (number int, first_name string);
load data inpath '.../doctors.avro' into table doctors;
alter table doctors add columns (last_name string);
load data inpath '../doctors_with_last_names' into table doctors;
select * from doctors; --should return both sets.
{noformat}
> Add support to ALTER TABLE .. ADD COLUMN to Avro backed tables
> --------------------------------------------------------------
>
> Key: HIVE-7446
> URL: https://issues.apache.org/jira/browse/HIVE-7446
> Project: Hive
> Issue Type: New Feature
> Reporter: Ashish Kumar Singh
> Assignee: Ashish Kumar Singh
> Attachments: HIVE-7446.patch
>
>
> HIVE-6806 adds native support for creating hive table stored as Avro. It
> would be good to add support to ALTER TABLE .. ADD COLUMN to Avro backed
> tables.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
