[
https://issues.apache.org/jira/browse/HIVE-894?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12829393#action_12829393
]
Carl Steinbach commented on HIVE-894:
-------------------------------------
* The patch introduces a lot of checkstyle violations.
* The class is named SortedArrayUtils, but it operates exclusively on Lists?
SortedListUtils would be more accurate.
* I think it would make more sense to create a SortedArrayList class that
extends ArrayList than to provide a Utils class.
** This can be implemented cheaply by extending TreeMap, such that (k,v) is
(value, value count).
* The SortedArrayUtils class references concrete types (ArrayList and Double)
instead of interfaces (List and Comparable).
** If you don't want to make it generic then I think this code should be rolled
back into the UDF classes since in its present form it doesn't earn the "Util"
moniker.
* It should be possible to create a common base class for Min and Max
implementations and share nearly all of the code. The only difference between
the two is the comparator.
> add udaf max_n, min_n to contrib
> --------------------------------
>
> Key: HIVE-894
> URL: https://issues.apache.org/jira/browse/HIVE-894
> Project: Hadoop Hive
> Issue Type: New Feature
> Affects Versions: 0.6.0
> Reporter: Zheng Shao
> Assignee: Zheng Shao
> Attachments: HIVE-894.1.patch, HIVE-894.1.patch, HIVE-894.2.patch
>
>
> These 2 UDAFs should return the max n numbers, min n numbers in order.
> {code}
> SELECT max_n(userid, 3) FROM src;
> [999,997,996]
> SELECT max_n(userid, 3)[1] FROM src;
> 997
> SELECT min_n(userid, 3) FROM src;
> [0, 3, 8]
> SELECT min_n(userid, 3)[2] FROM src;
> 8
> {code}
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
