You are viewing a plain text version of this content. The canonical link for it is here.
Posted to github@arrow.apache.org by "mustafasrepo (via GitHub)" <gi...@apache.org> on 2023/06/14 12:21:48 UTC
[GitHub] [arrow-datafusion] mustafasrepo opened a new pull request, #6665: Combine evaluate_stateful and evaluate_inside_range
mustafasrepo opened a new pull request, #6665:
URL: https://github.com/apache/arrow-datafusion/pull/6665
# 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 #.
# 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.
-->
# What changes are included in this PR?
This PR implements the proposal in the [#PR12](https://github.com/alamb/arrow-datafusion/pull/12)
- `evaluate_stateful` and `evaluate_inside_range` are combined under single `evaluate` methods
- `uses_window_frame` and `supports_bounded_execution` methods are moved from `BoundedWindowFunctionExpr` to `PartitionEvaluator`.
With the changes in this PR new Evaluators can be implemented according to table below.
|uses_window_frame|supports_bounded_execution|function_to_implement|
|---|---|----|
|false|false|`evaluate_all` (if we were to implement `PERCENT_RANK` it would end up in this quadrant, we cannot produce any result without seeing whole data)|
|false|true|`evaluate` (optionally can also implement `evaluate_all` for more optimized implementation. However, there will be default implementation that is suboptimal) . If we were to implement `ROW_NUMBER` it will end up in this quadrant. Example `OddRowNumber` showcases this use case|
|true|false|`evaluate` (I think as long as `uses_window_frame` is `true`. There is no way for `supports_bounded_execution` to be false). I couldn't come up with any example for this quadrant |
|true|true|`evaluate`. If we were to implement `FIRST_VALUE`, it would end up in this quadrant|.
<!--
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.
-->
# 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.
-->
<!--
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: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] mustafasrepo commented on a diff in pull request #6665: Combine evaluate_stateful and evaluate_inside_range
Posted by "mustafasrepo (via GitHub)" <gi...@apache.org>.
mustafasrepo commented on code in PR #6665:
URL: https://github.com/apache/arrow-datafusion/pull/6665#discussion_r1229645276
##########
datafusion/physical-expr/src/window/built_in_window_function_expr.rs:
##########
@@ -86,27 +86,10 @@ pub trait BuiltInWindowFunctionExpr: Send + Sync + std::fmt::Debug {
/// The default implementation does nothing
fn add_equal_orderings(&self, _builder: &mut OrderingEquivalenceBuilder) {}
- /// Can the window function be incrementally computed using
- /// bounded memory?
- ///
- /// If this function returns true, [`Self::create_evaluator`] must
- /// implement [`PartitionEvaluator::evaluate_stateful`]
- fn supports_bounded_execution(&self) -> bool {
- false
- }
-
- /// Does the window function use the values from its window frame?
- ///
- /// If this function returns true, [`Self::create_evaluator`] must
- /// implement [`PartitionEvaluator::evaluate_inside_range`]
- fn uses_window_frame(&self) -> bool {
- false
- }
-
/// Can this function be evaluated with (only) rank
///
/// If `include_rank` is true, then [`Self::create_evaluator`] must
- /// implement [`PartitionEvaluator::evaluate_with_rank`]
+ /// implement [`PartitionEvaluator::evaluate_with_rank_all`]
Review Comment:
For consistency, I have moved `include_rank` flag to `PartitionEvaluator` also.
--
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: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] alamb merged pull request #6665: Combine evaluate_stateful and evaluate_inside_range
Posted by "alamb (via GitHub)" <gi...@apache.org>.
alamb merged PR #6665:
URL: https://github.com/apache/arrow-datafusion/pull/6665
--
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: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] alamb commented on a diff in pull request #6665: Combine evaluate_stateful and evaluate_inside_range
Posted by "alamb (via GitHub)" <gi...@apache.org>.
alamb commented on code in PR #6665:
URL: https://github.com/apache/arrow-datafusion/pull/6665#discussion_r1230106114
##########
datafusion/physical-expr/src/window/rank.rs:
##########
@@ -115,12 +115,6 @@ pub(crate) struct RankEvaluator {
}
impl PartitionEvaluator for RankEvaluator {
- fn get_range(&self, idx: usize, _n_rows: usize) -> Result<Range<usize>> {
Review Comment:
❤️
--
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: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] alamb commented on a diff in pull request #6665: Combine evaluate_stateful and evaluate_inside_range
Posted by "alamb (via GitHub)" <gi...@apache.org>.
alamb commented on code in PR #6665:
URL: https://github.com/apache/arrow-datafusion/pull/6665#discussion_r1229589764
##########
datafusion/physical-expr/src/window/built_in_window_function_expr.rs:
##########
@@ -86,27 +86,10 @@ pub trait BuiltInWindowFunctionExpr: Send + Sync + std::fmt::Debug {
/// The default implementation does nothing
fn add_equal_orderings(&self, _builder: &mut OrderingEquivalenceBuilder) {}
- /// Can the window function be incrementally computed using
- /// bounded memory?
- ///
- /// If this function returns true, [`Self::create_evaluator`] must
- /// implement [`PartitionEvaluator::evaluate_stateful`]
- fn supports_bounded_execution(&self) -> bool {
- false
- }
-
- /// Does the window function use the values from its window frame?
- ///
- /// If this function returns true, [`Self::create_evaluator`] must
- /// implement [`PartitionEvaluator::evaluate_inside_range`]
- fn uses_window_frame(&self) -> bool {
- false
- }
-
/// Can this function be evaluated with (only) rank
///
/// If `include_rank` is true, then [`Self::create_evaluator`] must
- /// implement [`PartitionEvaluator::evaluate_with_rank`]
+ /// implement [`PartitionEvaluator::evaluate_with_rank_all`]
Review Comment:
Do you plan to give the same treatment (move to `PartitionEvaluator`) to `include_rank`?
##########
datafusion/physical-expr/src/window/partition_evaluator.rs:
##########
@@ -132,23 +140,40 @@ pub trait PartitionEvaluator: Debug + Send {
/// Called for window functions that *do not use* values from the
/// the window frame, such as `ROW_NUMBER`, `RANK`, `DENSE_RANK`,
- /// `PERCENT_RANK`, `CUME_DIST`, `LEAD`, `LAG`).
- fn evaluate(&self, _values: &[ArrayRef], _num_rows: usize) -> Result<ArrayRef> {
- Err(DataFusionError::NotImplemented(
- "evaluate is not implemented by default".into(),
- ))
+ /// `PERCENT_RANK`, `CUME_DIST`, `LEAD`, `LAG`). It produces result
+ /// of all rows in a single pass. It expects to receive whole table data
+ /// as a single batch.
+ fn evaluate_all(&mut self, values: &[ArrayRef], num_rows: usize) -> Result<ArrayRef> {
+ // When window frame boundaries are not used and evaluator supports bounded execution
+ // We can calculate evaluate result by repeatedly calling `self.evaluate` `num_rows` times
+ // If user wants to implement more efficient version, this method should be overwritten
+ // Default implementation may behave suboptimally (For instance `NumRowEvaluator` overwrites it)
+ if !self.uses_window_frame() && self.supports_bounded_execution() {
+ let res = (0..num_rows)
+ .map(|_idx| self.evaluate(values, &Range { start: 0, end: 1 }))
+ .collect::<Result<Vec<_>>>()?;
+ ScalarValue::iter_to_array(res.into_iter())
+ } else {
+ Err(DataFusionError::NotImplemented(
+ "evaluate_all is not implemented by default".into(),
Review Comment:
```suggestion
format!("evaluate_all is not implemented for {} when using window frames", self.name()),
```
I think the error could be more helpful, but I can also add that as a follow on PR
--
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: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] mustafasrepo commented on a diff in pull request #6665: Combine evaluate_stateful and evaluate_inside_range
Posted by "mustafasrepo (via GitHub)" <gi...@apache.org>.
mustafasrepo commented on code in PR #6665:
URL: https://github.com/apache/arrow-datafusion/pull/6665#discussion_r1229614204
##########
datafusion/physical-expr/src/window/built_in_window_function_expr.rs:
##########
@@ -86,27 +86,10 @@ pub trait BuiltInWindowFunctionExpr: Send + Sync + std::fmt::Debug {
/// The default implementation does nothing
fn add_equal_orderings(&self, _builder: &mut OrderingEquivalenceBuilder) {}
- /// Can the window function be incrementally computed using
- /// bounded memory?
- ///
- /// If this function returns true, [`Self::create_evaluator`] must
- /// implement [`PartitionEvaluator::evaluate_stateful`]
- fn supports_bounded_execution(&self) -> bool {
- false
- }
-
- /// Does the window function use the values from its window frame?
- ///
- /// If this function returns true, [`Self::create_evaluator`] must
- /// implement [`PartitionEvaluator::evaluate_inside_range`]
- fn uses_window_frame(&self) -> bool {
- false
- }
-
/// Can this function be evaluated with (only) rank
///
/// If `include_rank` is true, then [`Self::create_evaluator`] must
- /// implement [`PartitionEvaluator::evaluate_with_rank`]
+ /// implement [`PartitionEvaluator::evaluate_with_rank_all`]
Review Comment:
We can do so, what do you think about it. Should we do that?
--
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: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] mustafasrepo commented on a diff in pull request #6665: Combine evaluate_stateful and evaluate_inside_range
Posted by "mustafasrepo (via GitHub)" <gi...@apache.org>.
mustafasrepo commented on code in PR #6665:
URL: https://github.com/apache/arrow-datafusion/pull/6665#discussion_r1229614204
##########
datafusion/physical-expr/src/window/built_in_window_function_expr.rs:
##########
@@ -86,27 +86,10 @@ pub trait BuiltInWindowFunctionExpr: Send + Sync + std::fmt::Debug {
/// The default implementation does nothing
fn add_equal_orderings(&self, _builder: &mut OrderingEquivalenceBuilder) {}
- /// Can the window function be incrementally computed using
- /// bounded memory?
- ///
- /// If this function returns true, [`Self::create_evaluator`] must
- /// implement [`PartitionEvaluator::evaluate_stateful`]
- fn supports_bounded_execution(&self) -> bool {
- false
- }
-
- /// Does the window function use the values from its window frame?
- ///
- /// If this function returns true, [`Self::create_evaluator`] must
- /// implement [`PartitionEvaluator::evaluate_inside_range`]
- fn uses_window_frame(&self) -> bool {
- false
- }
-
/// Can this function be evaluated with (only) rank
///
/// If `include_rank` is true, then [`Self::create_evaluator`] must
- /// implement [`PartitionEvaluator::evaluate_with_rank`]
+ /// implement [`PartitionEvaluator::evaluate_with_rank_all`]
Review Comment:
We can do so, what do you think about it?
--
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: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org