[
https://issues.apache.org/jira/browse/HIVE-554?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12772815#action_12772815
]
Zheng Shao commented on HIVE-554:
---------------------------------
Sorry about the delay in reviewing.
* Can we reuse the same array/map in evaluate() functions? That will improve
the efficiency.
* For the display string, can we use "array(1,2,3)" format? Otherwise "[2]"
will be ambiguous. Maybe for map, we can also use "map(1:2, 'a':'b')".
> Add GenericUDF to create arrays, maps, and structs
> --------------------------------------------------
>
> Key: HIVE-554
> URL: https://issues.apache.org/jira/browse/HIVE-554
> Project: Hadoop Hive
> Issue Type: New Feature
> Reporter: Zheng Shao
> Assignee: Paul Yang
> Attachments: HIVE-554.1.patch
>
>
> Here is an example:
> {code}
> SELECT array(1,2,3)[3], map("a":1,"b":2,"c":3)["a"], struct(user_id:3,
> revenue: sum(rev))
> FROM table
> GROUP BY user_id;
> {code}
> This is relatively easy to do with the GenericUDF framework, and will greatly
> increase the flexibility of the language.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
