You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@calcite.apache.org by "Vladimir Sitnikov (JIRA)" <ji...@apache.org> on 2019/02/05 14:10:00 UTC

[jira] [Commented] (CALCITE-2301) JDBC adapter: use query timeout from the top-level statement

    [ https://issues.apache.org/jira/browse/CALCITE-2301?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16760822#comment-16760822 ] 

Vladimir Sitnikov commented on CALCITE-2301:
--------------------------------------------

{quote}A test for just PostgreSQL would be fine. I don't expect this to break often.{quote}
[~julianhyde], unfortunately, test for PostgreSQL it is not possible at the moment.

I guess {{select pg_sleep(20)}} should be enough to emulate long PostgreSQL query, however Calcite doesn't know {{pg_sleep}} function.
Then Calcite does not allow to customize OperatorTable (there are 3 fun=oracle,spartial,... hard-coded operator tables, and extensibility is not there).

I think PR#995 is good enough as it is an obvious improvement over previous hard-coded 10seconds (which did not have tests as well).

I'm committing PR#995 in case no-one objects within 24 hours.

> JDBC adapter: use query timeout from the top-level statement
> ------------------------------------------------------------
>
>                 Key: CALCITE-2301
>                 URL: https://issues.apache.org/jira/browse/CALCITE-2301
>             Project: Calcite
>          Issue Type: Bug
>          Components: core
>    Affects Versions: 1.16.0
>            Reporter: Fan Yang
>            Assignee: Julian Hyde
>            Priority: Minor
>              Labels: pull-request-available
>             Fix For: next
>
>
> It's not a good idea to have the magic number here. Also, databases may not get back within 10 second for various reasons (e.g., in the case of JDBC schema).



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