[
https://issues.apache.org/jira/browse/HIVE-6393?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13925315#comment-13925315
]
Lefty Leverenz commented on HIVE-6393:
--------------------------------------
Good, although I'd say "Hive" instead of "we" and mention the release number
(with a link to this jira). Or should it be "the query optimizer" instead of
"we"?
This can go in a version-info box at the end of the Join Syntax section, then a
simple example can be added just for emphasis at the end of the Examples
section. The syntax itself doesn't need any changes.
Your second sentence says that the column names can't be identical, so the
first example in the wikidoc can't use unqualified column references: "SELECT
a.* FROM a JOIN b ON (a.id = b.id)". But is it more restrictive than that --
neither table can have an (unreferenced) column named the same as the other
table's referenced column? Perhaps I'm overthinking this.
I can put this in the wiki after you fine-tune it, unless you'd rather do it
yourself.
> Support unqualified column references in Joining conditions
> -----------------------------------------------------------
>
> Key: HIVE-6393
> URL: https://issues.apache.org/jira/browse/HIVE-6393
> Project: Hive
> Issue Type: Improvement
> Reporter: Harish Butani
> Assignee: Harish Butani
> Fix For: 0.13.0
>
> Attachments: HIVE-6393.1.patch, HIVE-6393.2.patch, HIVE-6393.3.patch
>
>
> Support queries of the form:
> {noformat}
> create table r1(a int);
> create table r2(b);
> select a, b
> from r1 join r2 on a = b
> {noformat}
> This becomes more useful in old style syntax:
> {noformat}
> select a, b
> from r1, r2
> where a = b
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.2#6252)
