-
Notifications
You must be signed in to change notification settings - Fork 262
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
feat: lateral function call to linear data flow #536
Conversation
gaurav274
commented
Dec 16, 2022
- A simpler and easier-to-understand pipeline
- Reduce need for special case handling in the execution engine
- Simpler implementation for supporting reuse and ray due to the lack of branching.
def apply(self, before: LogicalJoin, context: OptimizerContext): | ||
# LogicalJoin(Lateral) LogicalApplyAndMerge | ||
# / \ -> | | ||
# A LogicalFunctionScan A |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Can we directly create the new operator (LogicalApplyAndMerge --- A) from the statement? I can see the current implementation requires few code modification. Does this transformation have any other benefits?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I considered doing it. However, I feel that will require adding a lot more core. Via rules, it is easier to understand the changes, and we can fallback to lateral join whenever needed. What do you think?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I can not think of any issues right now. We need to make sure that this rule applies first, because I assume all later rewrite rules (such as UDF reuse) will base on the new linear design.