[ 
https://issues.apache.org/jira/browse/LENS-88?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14264676#comment-14264676
 ] 

Suma Shivaprasad commented on LENS-88:
--------------------------------------

[~amareshwari] [~prongs] My earlier comment was wrt to

-> What are we planning to resolve statedim chain here? Shall we make this 
query valid only if there is one chain to reach statedim. If there is more than 
one chain, shall we fail the query to choose proper chain name instead? 

I think its better to choose one among the N join chains automatically instead 
of returning an error in this case since the end user may not be aware which 
joinchain he should choose and the system should be able to make a default 
decision with least no of hops as default or weights if available. If he needs 
to override default behaviour, he could choose a chain name instead.  I 
misinterpreted the comment earlier and just realised that throwing an error 
during join chain creation does not apply in this case. However I think we 
should go with the automatic resolution of chains in the system instead of 
throwing an error.



> Support two different chains for same in same query
> ---------------------------------------------------
>
>                 Key: LENS-88
>                 URL: https://issues.apache.org/jira/browse/LENS-88
>             Project: Apache Lens
>          Issue Type: Task
>          Components: cube
>            Reporter: Amareshwari Sriramadasu
>            Assignee: Rajat Khandelwal
>             Fix For: 2.0
>
>
> LENS-41 adds capability to specify multiple join chains and querying the 
> columns through chain.
> This task is to support two chains leading to same destination table with 
> different chains.



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

Reply via email to