You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Martijn Visser (Jira)" <ji...@apache.org> on 2022/06/20 14:26:00 UTC

[jira] [Comment Edited] (FLINK-28132) Should we remove type validations when using lookup-key join ?

    [ https://issues.apache.org/jira/browse/FLINK-28132?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17556404#comment-17556404 ] 

Martijn Visser edited comment on FLINK-28132 at 6/20/22 2:25 PM:
-----------------------------------------------------------------

But it's not only RDBMs; it's also Flink internally. I could see some danger. 

Either way, there was a FLIP drafted on implicit type coercion (https://cwiki.apache.org/confluence/display/FLINK/FLIP-154%3A+SQL+Implicit+Type+Coercion) but that discussion has never been completed and the FLIP hasn't been voted on. A change such as this one would definitely require this FLIP discussion to be restarted and voted on. 


was (Author: martijnvisser):
But it's not only RDBMs; it's also Flink internally. I could see some dangers. 

Either way, there was a FLIP drafted on implicit type coercion (https://cwiki.apache.org/confluence/display/FLINK/FLIP-154%3A+SQL+Implicit+Type+Coercion) but that discussion has never been completed and the FLIP hasn't been voted on. A change such as this one would definitely require this FLIP discussion to be restarted and voted on. 

> Should we remove type validations when using lookup-key join ?
> --------------------------------------------------------------
>
>                 Key: FLINK-28132
>                 URL: https://issues.apache.org/jira/browse/FLINK-28132
>             Project: Flink
>          Issue Type: Improvement
>          Components: Connectors / JDBC, Table SQL / Planner
>    Affects Versions: 1.15.0
>            Reporter: Zha Ji
>            Priority: Major
>
> As described in https://issues.apache.org/jira/browse/FLINK-18234
>  
> Execute sql
> {color:#6a8759}select * from t left join jdbc_source for system_time as of t.proctime AS j on t.id = j.id{color}
> t.id(VARCHAR) j.id(INT) will throw exception 
> org.apache.flink.table.api.TableException: VARCHAR(2147483647) and INTEGER does not have common type now
> If I remove some type validation codes, the sql works well on MySQL
>  
> Is it necessary to check data types when we join stream data to dynamic tables



--
This message was sent by Atlassian Jira
(v8.20.7#820007)