Improve benchmark performance

[andygrove]

Is your feature request related to a problem or challenge? Please describe what you are trying to do. Google document for a discussion of ideas to improve performance:

https://docs.google.com/document/d/16xFYLCzCcRQCk6UgHried-nodVF6NV7kFVpJEFV31eY/edit?usp=sharing

Describe the solution you'd like

Describe alternatives you've considered

Additional context

[mingmwang]

👍

[mingmwang]

I'm working on partition reasoning now which could help to avoid unnecessary shuffle.

https://github.com/apache/arrow-ballista/issues/284

[mingmwang]

If any issues/gaps are identified in the physical planning phase, you can assign those issue to me and I can take care.

[yahoNanJing]

👍