berkaysynnada opened a new pull request, #8653: URL: https://github.com/apache/arrow-datafusion/pull/8653
## Which issue does this PR close? <!-- We generally require a GitHub issue to be filed for all bug fixes and enhancements and this helps us generate change logs for our releases. You can link an issue to this PR using the GitHub syntax. For example `Closes #123` indicates that this PR will close issue #123. --> The continuation of the work initiated in https://github.com/apache/arrow-datafusion/pull/8395. It could also be beneficial for https://github.com/apache/arrow-datafusion/pull/7942. ## Rationale for this change <!-- Why are you proposing this change? If this is already explained clearly in the issue then this section is not needed. Explaining clearly why changes are proposed helps reviewers understand your changes and offer better suggestions for fixes. --> `TreeNode` implementations of some optimizer rules are challenging to understand and are open to misuse. This refactor standardizes the implementations and eliminates unnecessary payloads. ## What changes are included in this PR? <!-- There is no need to duplicate the description in the issue here but it is sometimes worth providing a summary of the individual changes in this PR. --> These implementations have been refactored: 1) DistributionContext 2) PlanWithCorrespondingSort 3) PlanWithCorrespondingCoalescePartitions 4) PipelineStatePropagator 5) OrderPreservationContext 6) SortPushDown 7) ExprOrdering `map_children()` functions of these implementations are now uniform. Previously, some of the rules were in `map_children()`, others were in some utils such as `new_from_children()`, and some were in transformer rules. This distributed structure made understanding and maintenance difficult. All these rules have now been moved into functions used as transform arguments on the optimizer part. Since Datafusion trees generally consist of nodes that store their children, each transform can implicitly have bottom-up transform capability. In some uses of `transform_up()`, after updating the children of the self node, additional logic is added during their attachment to the self node. This practice has been avoided. Now, all `map_children()` does is attach the updated children to the default-created self node without any modification of the self node. A similar situation may cause an algorithm that is expected to perform` transform_down()` to perform an implicit `transform_up()` if the `map_children()` implementation of the rule that perform `transform_down()` includes some logic. Perhaps a more comprehensive tree visitor-transformer design can be planned to address this issue. ## Are these changes tested? <!-- We typically require tests for all PRs in order to: 1. Prevent the code from being accidentally broken by subsequent changes 2. Serve as another way to document the expected behavior of the code If tests are not included in your PR, please explain why (for example, are they covered by existing tests)? --> Yes, with existing tests. ## Are there any user-facing changes? <!-- If there are user-facing changes then we may require documentation to be updated before approving the PR. --> <!-- If there are any breaking changes to public APIs, please add the `api change` label. --> -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
