zhen
zhen
it doesn't, because `temp` could easily be optimised into a register.
> > BTW, with CBO off, where do we use row count? > > we use it in places like : > > https://github.com/apache/spark/blob/161c596cafea9c235b5c918d8999c085401d73a9/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/statsEstimation/SizeInBytesOnlyStatsPlanVisitor.scala#L93-L100 > > where we just multiply...
sunchao commented the spark type is intended for maintaining type precision, but here the spark type is carrying the name `c1` which is renamed from `c0` so carrying it forward...
apart from that, what about SupportsMetadata interface in DSV2 base implementation?
nice addition.
could you exclude those style changes first (if it is indeed a must, do it in another JIRA). so that focus can be placed on the real changes intended. in...
quite some changes, you might want to squash your change into one commit. the styles and tests look good.
looks in pretty good shape.