Jefffrey opened a new pull request, #10023:
URL: https://github.com/apache/arrow-datafusion/pull/10023
## 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.
-->
Closes #9573
## 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.
-->
Rather than recursively constructing the initial physical plan from the top
down, which can lead to stack overflow errors, iteratively construct the plan
from the bottom up, which is also done concurrently.
## 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.
-->
Split up `create_initial_plan` to do the mapping from `LogicalPlan` to
`Arc<dyn ExecutionPlan` in a separate function for organization.
In `create_initial_plan` we first DFS the tree to get a flat Vec
representation, which stores `&LogicalPlan` to not require duplicating the tree
(this traversal is iterative).
With the flat tree, we can spawn async tasks from the leaves, which will
attempt to build the individual branches of the trees from the bottom up,
towards the root.
When these tasks encounter a node with 2 or more children, which represents
a collision point with other tasks, they append their current tree branch to
this parent node (which has it's current children branches which are ready
behind a `Mutex<Vec<_>>` for concurrent safety) and then check if there are
enough children to build the node. If not, this task terminates forever.
If there are enough children, this means the current task is the last node
to reach the parent, and can construct the parent and then continue to traverse
up the parent towards the root.
This continues until the number of tasks reduces to 1 and it emits the root
of the tree.
## 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)?
-->
## Are there any user-facing changes?
<!--
If there are user-facing changes then we may require documentation to be
updated before approving the PR.
-->
No
<!--
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]
