[Question] Skip Partial stage aggregation stage for sorted input

If the input is sorted by the group keys can't we just skip partial aggregation stage?

Review the aggregation planning and execution path in DataFusion, starting with how sorted group-key input is represented and how the partial aggregation stage is selected. Determine the correctness and performance criteria for skipping that stage; done means the issue has a decided design and corresponding implementation and validation plan.

Written by the indexing model from the issue text.