You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "blues zheng (JIRA)" <ji...@apache.org> on 2018/05/22 09:58:00 UTC

[jira] [Assigned] (FLINK-8379) Improve type checking for DataView

     [ https://issues.apache.org/jira/browse/FLINK-8379?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

blues zheng reassigned FLINK-8379:
----------------------------------

    Assignee: blues zheng

> Improve type checking for DataView
> ----------------------------------
>
>                 Key: FLINK-8379
>                 URL: https://issues.apache.org/jira/browse/FLINK-8379
>             Project: Flink
>          Issue Type: Improvement
>          Components: Table API &amp; SQL
>            Reporter: Timo Walther
>            Assignee: blues zheng
>            Priority: Major
>
> At the moment an accumulator with no proper type information is a valid accumulator.
> {code}
> 	public static class CountDistinctAccum {
> 		public MapView<Tuple, Integer> map;
> 		public long count;
> 	}
> {code}
> I quickly looked into the code and it seems that MapView with type information for key and value can be null. We should add a null check at the correct position to inform the user about the non-existing type information. We should also add the type check added with FLINK-8139 for the key type of MapView.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)