phoenix-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sokolov Yura (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (PHOENIX-2069) Support custom aggregate UDFs
Date Wed, 13 Dec 2017 09:55:00 GMT

    [ https://issues.apache.org/jira/browse/PHOENIX-2069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16289002#comment-16289002
] 

Sokolov Yura commented on PHOENIX-2069:
---------------------------------------

Vote for this feature.

My friend had to patch phoenix in order to add hyper-log-log aggregation. It was better to
make it aggregate udf instead, but there were no possibility.

> Support custom aggregate UDFs
> -----------------------------
>
>                 Key: PHOENIX-2069
>                 URL: https://issues.apache.org/jira/browse/PHOENIX-2069
>             Project: Phoenix
>          Issue Type: Bug
>            Reporter: James Taylor
>            Assignee: Rajeshbabu Chintaguntla
>              Labels: YARN-TLS
>
> We should build on our existing feature of custom UDFs to support custom aggregate functions.
We could drive the requirements initially from a YARN use case to allow them to plugin a custom
aggregator that would just be passed the Tuple (i.e. List<Cell> for each row scanned)
plus a context object. The custom code code then do whatever aggregation is required and we'd
have a hook to serialize back out the per scan aggregation.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Mime
View raw message