[
https://issues.apache.org/jira/browse/FLINK-1091?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Stephan Ewen updated FLINK-1091:
--------------------------------
Comment: was deleted
(was: The rational behind that decision was that, in order to access a
partitioned state, the accessing operator must be partitioned as well.
Making the access via a join (cogroup) assures that and outlined a consistent
way to integrate into the optimizer.
We are bit by bit adding operations with more explicit control, so explicit
access to the state would be a logical step. We need to think about whether to
mark an operator as "state accessing", or whether we should totally leave it to
the user to correctly partition the data (for example like in pull request
https://github.com/apache/incubator-flink/pull/108))
> Allow joins with the solution set using key selectors
> -----------------------------------------------------
>
> Key: FLINK-1091
> URL: https://issues.apache.org/jira/browse/FLINK-1091
> Project: Flink
> Issue Type: Improvement
> Components: Iterations
> Reporter: Vasia Kalavri
> Priority: Minor
> Labels: easyfix, features
>
> Currently, the solution set may only be joined with using tuple field
> positions.
> A possible solution can be providing explicit functions "joinWithSolution"
> and "coGroupWithSolution" to make sure the keys used are valid.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
