You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In subqueries, you might have relations as children of scalar operators. Although we are going to try to unnest the subquery in the optimizer, we should still be able to represent it in the logical plan and maybe also support directly executing the subquery without optimization.
The current IR assumes that scalar always has scalar has children, which is not the case when we add subquery support.
Suggestion
we can maybe have ScalarOperator<Scalar, Relation>, but only for the scalar operators that take in relation as children, you will have something like Subquery<Relation, Scalar>. For others, do what we have right now e.g. Add<Scalar>.
Resources
Unnesting Arbitrary Queries - Mark Raasveldt youtubeslides.
Problem
In subqueries, you might have relations as children of scalar operators. Although we are going to try to unnest the subquery in the optimizer, we should still be able to represent it in the logical plan and maybe also support directly executing the subquery without optimization.
The current IR assumes that scalar always has scalar has children, which is not the case when we add subquery support.
Suggestion
we can maybe have
ScalarOperator<Scalar, Relation>
, but only for the scalar operators that take in relation as children, you will have something likeSubquery<Relation, Scalar>
. For others, do what we have right now e.g.Add<Scalar>
.Resources
Unnesting Arbitrary Queries - Mark Raasveldt youtube slides.
Originally posted by @yliang412 in #4 (comment)
The text was updated successfully, but these errors were encountered: