[DISCUSS]Is this idea feasible?

[Andy <an...@163.com>]

————— 2022-5-27 —————

Xu Jie 10:30
I have an idea. SQL gets the table after routing through the engine, analyzes the table operation type, and if it is of DML or DDL type, set the task corresponding to the cache to cache invalidation. Do you think this idea is feasible?

Xu Jie 10:30
@peacewong@WDS 

peacewong@WDS 10:37
It may be necessary to maintain a large mapping, which tables are involved in each sql, and the idea feels no problem.

Xu Jie 10:46
Your current cache is based on JOB, is it not based on SQL? @peacewong@WDS 

peacewong@WDS 10:48
Now linkis provides based on job+time

Xu Jie 10:53
OK


—————  2022-5-27  —————

徐杰 10:30
我有个想法，SQL通过引擎路由后获取到表，分析表操作类型，如果是DML或者DDL类型的，就将cache对应的任务置成缓存无效，你觉得这个思路可行吗？

徐杰 10:30
@peacewong@WDS 

peacewong@WDS 10:37
可能要维护一个很大的映射，每个sql涉及了哪些表这些，思路感觉没问题的。

徐杰 10:46
你目前缓存是基于JOB吧，不是基于SQL的吧？@peacewong@WDS 

peacewong@WDS 10:48
现在linkis提供的是基于job+时间的

徐杰 10:53
好的

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@linkis.apache.org
For additional commands, e-mail: dev-help@linkis.apache.org