[
https://issues.apache.org/jira/browse/FLINK-10964?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16711232#comment-16711232
]
Timo Walther commented on FLINK-10964:
--------------------------------------
I can give more information here. The issue is rather a conceptual problem.
Currently, we only maintain results until a job has finished. If the job is
bounded (for batch or streaming with bounded sources), the result store is
cleaned up even though the CLI still wants to page through the result. A
solution would be:
- keep results in {{ResultStore}} even if query has been stopped
- delete result when {{Executor#cancelQuery}} is called
- or by a timeout that has been specified using the "execution" properties
(meaning we need a timer service in the result store).
> sql-client throws exception when paging through finished batch query
> ---------------------------------------------------------------------
>
> Key: FLINK-10964
> URL: https://issues.apache.org/jira/browse/FLINK-10964
> Project: Flink
> Issue Type: Bug
> Components: SQL Client
> Reporter: Seth Wiesman
> Assignee: vinoyang
> Priority: Major
>
> When paging through a batch query in state 'Finished' the sql client throws
> the following exception:
> {code:java}
> org.apache.flink.table.client.gateway.SqlExecutionException: Could not find a
> result with result identifier '0c7dce30d287fdd13b934fbefe5a38d1'.{code}
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
