[jira] [Resolved] (IMPALA-2569) Support schema evolution by omitting columns from insert

["Aman Sinha (Jira)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/IMPALA-2569?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Aman Sinha resolved IMPALA-2569.
--------------------------------
    Resolution: Later

Inserting nulls implicitly for the additional columns could lead to some unexpected behavior.  The 'DEFAULT' keyword is the preferred approach that some RDBMS's (e.g Postgres: https://www.postgresql.org/docs/9.3/ddl-default.html)  have supported for quite a while since it puts the responsibility on the user to define a proper default value for the missing column value.  Also, Kudu already supports this DEFAULT keyword for columns (see https://docs.cloudera.com/documentation/enterprise/latest/topics/impala_kudu.html#kudu_default_attribute). We may want to think about similar extension to the Create table ddl.  I am resolving this JIRA for 'later'.

> Support schema evolution by omitting columns from insert
> --------------------------------------------------------
>
>                 Key: IMPALA-2569
>                 URL: https://issues.apache.org/jira/browse/IMPALA-2569
>             Project: IMPALA
>          Issue Type: Improvement
>          Components: Frontend
>    Affects Versions: Impala 2.2.4
>            Reporter: Peter Ebert
>            Priority: Minor
>
> Adding a column to a table will break existing sql that inserts using 
> alter table table_name add columns (col3)
> insert into table_name
> select col1, col2 -- error thrown, expects 3 columns
> The workaround is to specify each column name in the insert, but many clients have 200-1000 column tables (Banks) where a select * would save them much development time when evolving their schema.  The default behavior should be as it is elsewhere, if the column is omitted nulls are inserted.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)