[
https://issues.apache.org/jira/browse/FLINK-24861?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17442186#comment-17442186
]
Wenlong Lyu commented on FLINK-24861:
-------------------------------------
hi, [~gaurav726] I have a different point: it is ok to cache the empty key by
default. The cache is usually used to avoid too frequent access on database,
empty result query should also be in the scope.
I also agrees that we can exclude empty result in case to improve the join
result when the dim table is increasing by time but not updating, so I think
you could add an optional config to exclude empty query result in cache
instead. what do you think?
> Flink MySQL Look cache update for empty hit
> -------------------------------------------
>
> Key: FLINK-24861
> URL: https://issues.apache.org/jira/browse/FLINK-24861
> Project: Flink
> Issue Type: New Feature
> Components: Connectors / JDBC
> Affects Versions: 1.14.0
> Reporter: Gaurav Miglani
> Assignee: Gaurav Miglani
> Priority: Major
> Labels: pull-request-available
>
> Ideally, in case of cache miss for a key, or with null value fetch for key,
> key shouldn't be cached
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
