[jira] [Commented] (FLINK-11039) LogicalExchange and HashPartitioner realization

[Fabian Hueske (JIRA)]

    [ 
https://issues.apache.org/jira/browse/FLINK-11039?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16708459#comment-16708459
 ] 

Fabian Hueske commented on FLINK-11039:
---------------------------------------

Please correct me, if I'm wrong, but this is only relevant if we do physical 
planning?
Right now, we manually convert the logical plan into a DataSet/DataStream 
program, i.e., physical translation happens (to some extend) in the 
{{translateToPlan()}} method. We won't see LogicalExchange nodes during 
optimization.

So, is this part of a larger effort to move to physical optimization? If so, 
this should be thoroughly discussed, IMO.

Best, Fabian

> LogicalExchange and HashPartitioner realization
> -----------------------------------------------
>
>                 Key: FLINK-11039
>                 URL: https://issues.apache.org/jira/browse/FLINK-11039
>             Project: Flink
>          Issue Type: Improvement
>          Components: Table API & SQL
>    Affects Versions: 1.6.2, 1.7.0
>            Reporter: ambition
>            Priority: Minor
>             Fix For: 1.7.0
>
>
> RelTimeIndicatorConverter.visit(exchange: LogicalExchange) function 
> realization.
> FlinkLogicalExchange realization 
> org.apache.calcite.rel.logical.LogicalExchange.
> HashPartitioner is Partitioner that implements hash-based partitioning using 
> Java's `Object.hashCode`. support org.apache.calcite.rel.RelDistribution.Type 
> HASH_DISTRIBUTED



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