[SPARK-38832][SQL][FOLLOWUP] Support propagate empty expression set for distinct key

[ulysses-you]

What changes were proposed in this pull request?

• Improve DistinctKeyVisitor that support propagate empty set
• Small improvement for match alias

Why are the changes needed?

Make distinct keys can be used to optimize more case, see comment #36117 (comment)

Does this PR introduce any user-facing change?

Improve performance

How was this patch tested?

add test

[ulysses-you]

cc @cloud-fan@sigmod@wangyum

[wangyum]

Could we filter out empty distinctKey for safe?

spark/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/Optimizer.scala

Lines 442 to 443 in 21e48b7

	agg.child.distinctKeys.exists(
	_.subsetOf(ExpressionSet(ae.aggregateFunction.children.filterNot(_.foldable)))) =>

[cloud-fan]

@wangyum what's wrong with that code?

[wangyum]

For example:

Set[ExpressionSet](ExpressionSet()).exists(_.subsetOf(ExpressionSet(ae.aggregateFunction.children.filterNot(_.foldable))))

It is always true, it may have potential problems.

[ulysses-you]

if we get an unexpected empty ExpressionSet it should be a bug .. shall we trust the framework ?

[cloud-fan]

I don't see why empty expression set is special. If the framework has bugs and produces incorrect distinct keys, we will have query correctness bugs.

[wangyum]

I'm ok if you do not think it is needed.

[cloud-fan]

My point is empty expression set is informative and we shouldn't skip it. It means the data is distinct by Nil, so it has at most 1 row and can override any other distinct keys.

[wangyum]

LGTM

[cloud-fan]

thanks, merging to master!

[ulysses-you]

thank you @wangyum@cloud-fan