fix: from_plan generate incorrect schema when it is Aggregate

[YjyJeff]

Rationale for this change

For the from_plan function, we want to generate a new logical plan with inputs and expressions replaced. When we encounter the Aggregate variant, the old code simply clones the old schema. However, the new group by expressions may have a different schema from the old one. Therefore, we should use the try_new function to create a new Aggregate variant with the correct schema.

Are these changes tested?

Yes

Are there any user-facing changes?

No

[jackwener]

There is related issue or specific example?

It's legacy code, exists with some reason. If there are no specific reasons, we should not change it.

[YjyJeff]

I have not created the issue, should we create an issue to talk about it?

I found this problem in our custom optimizer. In our optimizer, we will change the return type of the expression. For example, if we have the query

select col_0 / 3 as a, count(1) from table group by a 

and the type of col_0 is DataType::UInt32. Without our custom optimizer, the expression col_0 / 3 will return the type DataType::Int64（because the literal 3 is parsed as Int64）. Our optimizer will convert the literal Int64(3) to UInt(32) such that we can avoid the cost of casting col_0 to Int64Array. In this case, the return type of the group by expression changed from DataType::Int64 to DataType::UInt32.

It's legacy code, exists with some reason. If there are no specific reasons, we should not change it.

What are the reasons? In my view, from_plan may change the schema, therefore, the legacy code is a bug here. My patch fixed the bug and have passed all of the tests

[jackwener]

related PR: #6827 which revert #6595 #6820

related issue: #6819 #6790

[jackwener]

I found this problem in our custom optimizer. In our optimizer, we will change the return type of the expression. For example, if we have the query

currently, we can't change schema in optimization, because it will cause scheme match failed.

[YjyJeff]

Ok, I know the reasons now. However, the from_plan function looks pretty odd now: some logical plans support changing the schema but others do not. I think it still should be solved 0.0

[jackwener]

Ok, I know the reasons now. However, the from_plan function looks pretty odd now: some logical plans support changing the schema but others do not. I think it still should be solved 0.0

Yes, I think we indeed need to resolve this problem, but it's troublesome.

It needs to take more time to discuss how to solve this problem.

I add a comment #6790 (comment) to explain reason for this. Hope it can help others in community.

[jackwener]

cc @alamb