flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Chesnay Schepler (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (FLINK-3119) Remove dependency on Tuple from HadoopOutputFormat
Date Fri, 04 Dec 2015 18:43:11 GMT

    [ https://issues.apache.org/jira/browse/FLINK-3119?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15041953#comment-15041953
] 

Chesnay Schepler commented on FLINK-3119:
-----------------------------------------

if you want another serialization method could you not simply extend HadoopInputFormatBase,
which is not hardwired for tuples and contains 99% of the OutputFormat's functionality?

> Remove dependency on Tuple from HadoopOutputFormat
> --------------------------------------------------
>
>                 Key: FLINK-3119
>                 URL: https://issues.apache.org/jira/browse/FLINK-3119
>             Project: Flink
>          Issue Type: Improvement
>            Reporter: Nick Dimiduk
>            Priority: Minor
>
> {{HadoopOutputFormat}}'s dependency on {{Tuple}} precludes it's use by folks using other
serialization methods (ie, their own POJO's). Would be good to make this class more broadly
useful to folks, or somehow mark this implementation as specifically dependent on {{Tuple}}
-- it's not at all obvious from its name.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message