2010YOUY01 commented on PR #23738:
URL: https://github.com/apache/datafusion/pull/23738#issuecomment-5140852972

   > ```
   > #23828 physical operator
   > └── #23829 logical semantics and physical planning
   >     ├── #23830 SQL frontend and unparser
   >     │   └── #23833 benchmarks
   >     ├── #23831 DataFrame API
   >     └── #23832 protobuf serialization
   > ```
   
   I think this split is a good idea.
   
   One issue is that I think it's better to use `sqllogictest` as the major 
test coverage, since they're stronger and easier to maintain, comparing with 
UTs on only the `ExecutionPlan`; Also we might need the SQL benchmark to guide 
and validate the design for physical operator.
   
   So I think it's likely we have to 1. first implement a simple v1 physical 
operator 2. Implement SQL interface, then add more tests and benchmark through 
SQL interface, and potentially do some fixes/changes in step 1.
   
   But in order to make progress on such a large feature, it's still a better 
choice I guess 🤔 
   
   I plan to continue reviewing #23828 (step1), and try to get it merged first.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to