[ 
https://issues.apache.org/jira/browse/FLINK-7217?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Flink Jira Bot updated FLINK-7217:
----------------------------------
      Labels: auto-deprioritized-major auto-deprioritized-minor  (was: 
auto-deprioritized-major stale-minor)
    Priority: Not a Priority  (was: Minor)

This issue was labeled "stale-minor" 7 days ago and has not received any 
updates so it is being deprioritized. If this ticket is actually Minor, please 
raise the priority and ask a committer to assign you the issue or revive the 
public discussion.


> Give hints about unsupported SQL / Table API features.
> ------------------------------------------------------
>
>                 Key: FLINK-7217
>                 URL: https://issues.apache.org/jira/browse/FLINK-7217
>             Project: Flink
>          Issue Type: Improvement
>          Components: Table SQL / API
>            Reporter: Fabian Hueske
>            Priority: Not a Priority
>              Labels: auto-deprioritized-major, auto-deprioritized-minor
>
> The Table API and SQL do not support every type of query. While being valid 
> SQL, we cannot translate certain queries into execution plans. Especially, 
> the supported syntax of streaming queries is limited and will probably always 
> be limited.
> Therefore it is important to explain a user why a query could not be 
> translated. The problem here is that we need to wait for the optimizer to 
> explore the full search space and apply all transformation rules before we 
> can tell the user that the query could not be translated. Hence, we cannot 
> simply throw an exception if a rule cannot be applied, because another rule 
> might succeed. If we throw an exception in the optimizer, we would kill the 
> exploration and miss a valid plan. However, after the optimizer failed, we 
> are missing the complete context and cannot provide valuable feedback beyond 
> "query uses an unsupported feature".
> We should try to improve the situation and provide better feedback to users. 
> One approach would be that rules collect possible reasons why they could not 
> be applied and these reasons are returned to the user when the translation 
> fails.



--
This message was sent by Atlassian Jira
(v8.20.1#820001)

Reply via email to