You are viewing a plain text version of this content. The canonical link for it is here.
Posted to github@datafusion.apache.org by "2010YOUY01 (via GitHub)" <gi...@apache.org> on 2026/08/11 09:59:33 UTC
[PR] refactor: migrate ordered single aggregation [datafusion]
2010YOUY01 opened a new pull request, #24259:
URL: https://github.com/apache/datafusion/pull/24259
## 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.
-->
part of https://github.com/apache/datafusion/issues/22710
## 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.
Please explain the problem you are trying to solve in terms of the user-visible
behavior, rather than the implementation.
For example, "The code in `foo.rs` doesn't handle nulls" is a symptom of the
implementation. "COUNT(DISTINCT) returns wrong results when the column contains
nulls" is the user-visible problem.
-->
Migrate ordered single aggregation path.
For the background of the refactor, see EPIC issue for details
For the overview of the refactored path, see comments at the top of `datafusion/physical-plan/src/aggregates/ordered_single_stream.rs`
## 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.
-->
- `datafusion/physical-plan/src/aggregates/mod.rs`: physical planning changes that route the applicable execution to the 'ordered-single-aggregate` path
- `datafusion/physical-plan/src/aggregates/ordered_single_stream.rs`: main implementation. Overview is at the comments for `OrderedSingleAggregateStream`, and the entry point for execution state machine is at `poll_next()`
## 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)?
-->
Existing tests. I have verified the test line coverage from `llvm-cov`
## 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.
-->
No
--
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@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] refactor(hash-aggr): migrate ordered single aggregation [datafusion]
Posted by "alamb (via GitHub)" <gi...@apache.org>.
alamb commented on code in PR #24259:
URL: https://github.com/apache/datafusion/pull/24259#discussion_r3761852773
##########
datafusion/physical-plan/src/aggregates/mod.rs:
##########
@@ -1215,6 +1219,12 @@ impl AggregateExec {
)?));
}
+ if self.should_use_ordered_single_aggregate_stream(context) {
Review Comment:
FWIW while I like the explicit name of these methods I do wonder if the difference between them might be clearer if there was one giant match statement (rather than a bunch of if ... chains)
##########
datafusion/physical-plan/src/aggregates/mod.rs:
##########
@@ -672,12 +675,15 @@ enum StreamType {
OrderedPartialAggregate(OrderedPartialAggregateStream),
/// Final stage of aggregation for ordered input.
OrderedFinalAggregate(OrderedFinalAggregateStream),
+ /// Single stage of aggregation for ordered input.
+ OrderedSingleAggregate(OrderedSingleAggregateStream),
/// Hash aggregation reused for multiple stages
///
/// Note this is being incrementally migrated to dedicated streams like
/// [`StreamType::PartialHash`], [`StreamType::FinalHash`],
/// [`StreamType::OrderedPartialAggregate`], and
- /// [`StreamType::OrderedFinalAggregate`]
+ /// [`StreamType::OrderedFinalAggregate`], and
Review Comment:
Doc nit: this list now has a double "and" ("… and `OrderedFinalAggregate`, and `OrderedSingleAggregate`").
```suggestion
/// [`StreamType::OrderedPartialAggregate`],
/// [`StreamType::OrderedFinalAggregate`], and
```
##########
datafusion/physical-plan/src/aggregates/ordered_single_stream.rs:
##########
@@ -0,0 +1,886 @@
+// Licensed to the Apache Software Foundation (ASF) under one
+// or more contributor license agreements. See the NOTICE file
+// distributed with this work for additional information
+// regarding copyright ownership. The ASF licenses this file
+// to you under the Apache License, Version 2.0 (the
+// "License"); you may not use this file except in compliance
+// with the License. You may obtain a copy of the License at
+//
+// http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing,
+// software distributed under the License is distributed on an
+// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, either express or implied. See the License for the
+// specific language governing permissions and limitations
+// under the License.
+
+//! Single-stage aggregate stream for ordered raw input.
+
+use std::ops::ControlFlow;
+use std::sync::Arc;
+use std::task::{Context, Poll};
+
+use arrow::datatypes::SchemaRef;
+use arrow::record_batch::RecordBatch;
+use datafusion_common::{DataFusionError, Result, internal_datafusion_err, internal_err};
+use datafusion_execution::TaskContext;
+use datafusion_execution::memory_pool::{MemoryConsumer, MemoryReservation};
+use datafusion_physical_expr::PhysicalSortExpr;
+use datafusion_physical_expr::expressions::Column;
+use datafusion_physical_expr_common::sort_expr::LexOrdering;
+use futures::stream::{Stream, StreamExt};
+
+use super::aggregate_hash_table::{OrderedAggregateTable, SingleMarker};
+use super::group_values::GroupByMetrics;
+use super::ordered_final_stream::OrderedFinalAggregateStream;
+use super::{AggregateExec, create_schema};
+use crate::aggregates::AggregateMode;
+use crate::metrics::{BaselineMetrics, RecordOutput, SpillMetrics};
+use crate::sorts::IncrementalSortIterator;
+use crate::sorts::streaming_merge::{SortedSpillFile, StreamingMergeBuilder};
+use crate::spill::spill_manager::SpillManager;
+use crate::stream::EmptyRecordBatchStream;
+use crate::{InputOrderMode, RecordBatchStream, SendableRecordBatchStream};
+
+/// Single aggregate stream for `InputOrderMode::Sorted` and
+/// `InputOrderMode::PartiallySorted`.
+///
+/// # Example
+///
+/// SELECT k, AVG(v) FROM t GROUP BY k;
+///
+/// If the input is ordered by `k`, and there are existing key partitioning on group
+/// by keys, the single mode aggregation with ordering optimization can be used:
+///
+/// ## Plan
+/// AggregateExec(stage=single, ordered)
+/// -- DataSourceExec(t)
+///
+/// ## Single Stage Behavior
+/// Input: raw rows
+/// Output: final results for all groups (for example, `AVG(x)`)
+///
+/// # Order-based Optimization
+///
+/// For the aggregation work, the hash aggregation implementation is reused.
+///
+/// After each input batch, check whether any groups can be emitted eagerly to
+/// improve memory efficiency. For example, if the last group key seen is
+/// `k = 100`, it is safe to emit all groups with keys less than 100 because the
+/// input is ordered.
+///
+/// # Memory Pressure and Spilling
+///
+/// ## Fully ordered case
+///
+/// If the input is ordered by every group key, for example:
+///
+/// - Input order: `a, b`
+/// - `GROUP BY`: `a, b`
+///
+/// Completed groups can be emitted as soon as the next group is observed. Thus,
+/// only the current group remains active after completed groups are emitted, and
+/// memory usage does not grow with the total number of groups.
+///
+/// If a memory reservation nevertheless fails, the stream returns the error
+/// directly, indicating an unexpected behavior.
+///
+/// ## Partially ordered case
+///
+/// If the input is ordered by only a subset of the group keys, for example:
+///
+/// - Input order: `a`
+/// - `GROUP BY`: `a, b`
+///
+/// If one `a` value contains many distinct `b` values, the table may accumulate
+/// enough groups to exceed the memory limit.
+///
+/// On reservation failure, the stream sorts the current intermediate states by
+/// the complete group key and spills them as one run. After the input ends, it
+/// spills any remaining states, performs a sort-preserving merge of all runs,
+/// and feeds the merged input into a fully ordered final aggregate stream.
+pub(crate) struct OrderedSingleAggregateStream {
+ schema: SchemaRef,
+ input: SendableRecordBatchStream,
+ reservation: MemoryReservation,
+ baseline_metrics: BaselineMetrics,
+ state: Option<OrderedSingleAggregateState>,
+}
+
+/// Spill configuration and accumulated runs for partially ordered single
+/// aggregation.
+///
+/// Each spill event drains all currently buffered groups, sorts their intermediate
+/// states by the full group key, and writes them to one spill file. All files are
+/// merged and replayed after the original input ends.
+struct OrderedSingleSpillContext {
+ /// Aggregate configuration used to construct the final replay stream.
+ final_agg: AggregateExec,
+ /// Task context
+ context: Arc<TaskContext>,
+ /// Original partition index
+ partition: usize,
+ /// Target batch size from configuration
+ batch_size: usize,
+ /// Full group-key ordering, such ordering with be kept in: a) individual spill
+ /// files, b) order after final merging and streaming aggregate
+ spill_expr: LexOrdering,
+ /// Spill I/O and metrics manager.
+ spill_manager: SpillManager,
+ /// Fully sorted spill runs waiting to be merged.
+ spills: Vec<SortedSpillFile>,
+}
+
+/// See comments at `poll_next()` for details.
+enum OrderedSingleAggregateState {
+ ReadingInput {
+ table: OrderedAggregateTable<SingleMarker>,
+ /// None if either
+ /// - Disk Manager doesn't enable temporary file creation
+ /// - The group keys are fully ordered, it's expected to use bounded memory
+ spill_context: Option<Box<OrderedSingleSpillContext>>,
+ },
+ Spilling {
+ table: OrderedAggregateTable<SingleMarker>,
+ spill_context: Box<OrderedSingleSpillContext>,
+ },
+ ProducingOutput {
+ table: OrderedAggregateTable<SingleMarker>,
+ },
+ PreparingMergeInput {
+ table: OrderedAggregateTable<SingleMarker>,
+ spill_context: Box<OrderedSingleSpillContext>,
+ },
+ MergingSpills {
+ stream: SendableRecordBatchStream,
+ },
+ Done,
+ /// Sentinel state to use when returning error from any other states, because:
+ /// - It explicitly releases state-owned resources immediately
+ /// - More defensive against accidentally resuming execution after error
+ Error,
+}
+
+type OrderedSingleAggregatePoll = Poll<Option<Result<RecordBatch>>>;
+type OrderedSingleAggregateStateTransition = ControlFlow<
+ (OrderedSingleAggregatePoll, OrderedSingleAggregateState),
+ OrderedSingleAggregateState,
+>;
+
+impl OrderedSingleSpillContext {
+ fn new(
+ agg: &AggregateExec,
+ context: &Arc<TaskContext>,
+ partition: usize,
+ batch_size: usize,
+ input_order_mode: &InputOrderMode,
+ spill_schema: &SchemaRef,
+ spill_metrics: SpillMetrics,
+ ) -> Result<Self> {
+ let group_schema = agg.group_by.group_schema(&agg.input().schema())?;
+ let output_ordering = agg.cache.output_ordering();
+ let InputOrderMode::PartiallySorted(order_indices) = input_order_mode else {
+ return internal_err!(
+ "Ordered single spill requires partially ordered input"
+ );
+ };
+ let spill_indices = order_indices.iter().copied().chain(
+ (0..group_schema.fields().len()).filter(|idx| !order_indices.contains(idx)),
+ );
+ let spill_sort_exprs = spill_indices.map(|idx| {
+ let field = group_schema.field(idx);
+ let output_expr = Column::new(field.name(), idx);
+ let sort_options = output_ordering
+ .and_then(|ordering| ordering.get_sort_options(&output_expr))
+ .unwrap_or_default();
+ PhysicalSortExpr::new(Arc::new(output_expr), sort_options)
+ });
+ let Some(spill_expr) = LexOrdering::new(spill_sort_exprs) else {
+ return internal_err!("Ordered single spill expression is empty");
+ };
+
+ let spill_manager = SpillManager::new(
+ context.runtime_env(),
+ spill_metrics,
+ Arc::clone(spill_schema),
+ )
+ .with_compression_type(context.session_config().spill_compression());
+
+ // Spilled rows contain group keys and intermediate states. Replay must
+ // merge those states and evaluate the final aggregate values.
+ let mut final_agg = agg.clone();
+ final_agg.mode = match agg.mode {
+ AggregateMode::Single => AggregateMode::Final,
+ AggregateMode::SinglePartitioned => AggregateMode::FinalPartitioned,
+ mode => {
+ return internal_err!(
+ "Ordered single aggregate spill cannot replay aggregate mode {mode:?}"
+ );
+ }
+ };
+ final_agg.group_by = Arc::new(agg.group_by.as_final());
+ final_agg.input_order_mode = InputOrderMode::Sorted;
+
+ Ok(Self {
+ final_agg,
+ context: Arc::clone(context),
+ partition,
+ batch_size,
+ spill_expr,
+ spill_manager,
+ spills: vec![],
+ })
+ }
+
+ fn has_spills(&self) -> bool {
+ !self.spills.is_empty()
+ }
+
+ /// Sorts and spills the aggregated groups. Memory reservation should be updated
+ /// by the caller.
+ ///
+ /// Individual spill files are ordered by the `group by` keys.
+ ///
+ /// See [`OrderedSingleAggregateStream`] for spilling details.
+ fn spill_table(
+ &mut self,
+ table: &mut OrderedAggregateTable<SingleMarker>,
+ ) -> Result<()> {
+ let Some(batch) = table.take_state_batch()? else {
+ return Ok(());
+ };
+
+ let sorted_iter =
+ IncrementalSortIterator::new(batch, self.spill_expr.clone(), self.batch_size);
+ let spill_file = self
+ .spill_manager
+ .spill_record_batch_iter_and_return_max_batch_memory(
+ sorted_iter,
+ "OrderedSingleAggregateSpill",
+ )?;
+
+ let Some((file, max_record_batch_memory)) = spill_file else {
+ return internal_err!("Ordered single aggregation produced an empty spill");
+ };
+
+ self.spills.push(SortedSpillFile {
+ file,
+ max_record_batch_memory,
+ });
+
+ Ok(())
+ }
+
+ /// Merges every sorted run and finalizes it through the fully ordered path.
+ fn into_replay_stream(
Review Comment:
maybe worth a ticket to track
##########
datafusion/physical-plan/src/aggregates/mod.rs:
##########
@@ -1215,6 +1219,12 @@ impl AggregateExec {
)?));
}
+ if self.should_use_ordered_single_aggregate_stream(context) {
Review Comment:
For example the only difference between `should_use_single_hash_stream` and `should_use_ordered_single_aggregate_stream` seems to be the condition
```rust
&& self.input_order_mode != InputOrderMode::Linear
```
🤔
--
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@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] refactor(hash-aggr): migrate ordered single aggregation [datafusion]
Posted by "2010YOUY01 (via GitHub)" <gi...@apache.org>.
2010YOUY01 commented on code in PR #24259:
URL: https://github.com/apache/datafusion/pull/24259#discussion_r3757239487
##########
datafusion/physical-plan/src/aggregates/aggregate_hash_table/common_ordered.rs:
##########
@@ -304,18 +313,18 @@ impl<AggrMode> OrderedAggregateTable<AggrMode> {
EmitTo::All => (EmitTo::First(self.batch_size), false),
}
}
- /// Aggregates one evaluated input batch.
- ///
- /// This common utility is used by ordered partial and ordered final aggregation.
- ///
- /// # Argument: `is_final`
+
+ /// Aggregates one evaluated input batch after selecting the mode-specific
+ /// accumulator operation.
///
- /// - `true`: merge partial aggregate states for final aggregation.
- /// - `false`: update aggregate states from raw input for partial aggregation.
+ /// Each aggregation mode chooses a different `aggregate_fn` according to its
+ /// semantics. For example, partial aggregation takes raw inputs and updates
+ /// stored partial states, so it uses
+ /// [`datafusion_expr::GroupsAccumulator::update_batch`].
pub(super) fn aggregate_evaluated_batch(
&mut self,
evaluated_batch: &EvaluatedAggregateBatch,
- is_final: bool,
+ aggregate_fn: AggregateBatchFn,
Review Comment:
here is the key change for this file, just a clean-up to make it easier to be reused by different aggregation modes.
--
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@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] refactor(hash-aggr): migrate ordered single aggregation [datafusion]
Posted by "2010YOUY01 (via GitHub)" <gi...@apache.org>.
2010YOUY01 commented on code in PR #24259:
URL: https://github.com/apache/datafusion/pull/24259#discussion_r3772020445
##########
datafusion/physical-plan/src/aggregates/mod.rs:
##########
@@ -1215,6 +1219,12 @@ impl AggregateExec {
)?));
}
+ if self.should_use_ordered_single_aggregate_stream(context) {
Review Comment:
This is a good point, I've noticed it also. This entire code block can likely be simplified somehow to be more understandable.
I've opened a issue https://github.com/apache/datafusion/issues/24318, and I'll try to do it after the refactor is done.
--
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@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] refactor(hash-aggr): migrate ordered single aggregation [datafusion]
Posted by "2010YOUY01 (via GitHub)" <gi...@apache.org>.
2010YOUY01 commented on code in PR #24259:
URL: https://github.com/apache/datafusion/pull/24259#discussion_r3757233396
##########
datafusion/physical-plan/src/aggregates/ordered_single_stream.rs:
##########
@@ -0,0 +1,886 @@
+// Licensed to the Apache Software Foundation (ASF) under one
+// or more contributor license agreements. See the NOTICE file
+// distributed with this work for additional information
+// regarding copyright ownership. The ASF licenses this file
+// to you under the Apache License, Version 2.0 (the
+// "License"); you may not use this file except in compliance
+// with the License. You may obtain a copy of the License at
+//
+// http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing,
+// software distributed under the License is distributed on an
+// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, either express or implied. See the License for the
+// specific language governing permissions and limitations
+// under the License.
+
+//! Single-stage aggregate stream for ordered raw input.
+
+use std::ops::ControlFlow;
+use std::sync::Arc;
+use std::task::{Context, Poll};
+
+use arrow::datatypes::SchemaRef;
+use arrow::record_batch::RecordBatch;
+use datafusion_common::{DataFusionError, Result, internal_datafusion_err, internal_err};
+use datafusion_execution::TaskContext;
+use datafusion_execution::memory_pool::{MemoryConsumer, MemoryReservation};
+use datafusion_physical_expr::PhysicalSortExpr;
+use datafusion_physical_expr::expressions::Column;
+use datafusion_physical_expr_common::sort_expr::LexOrdering;
+use futures::stream::{Stream, StreamExt};
+
+use super::aggregate_hash_table::{OrderedAggregateTable, SingleMarker};
+use super::group_values::GroupByMetrics;
+use super::ordered_final_stream::OrderedFinalAggregateStream;
+use super::{AggregateExec, create_schema};
+use crate::aggregates::AggregateMode;
+use crate::metrics::{BaselineMetrics, RecordOutput, SpillMetrics};
+use crate::sorts::IncrementalSortIterator;
+use crate::sorts::streaming_merge::{SortedSpillFile, StreamingMergeBuilder};
+use crate::spill::spill_manager::SpillManager;
+use crate::stream::EmptyRecordBatchStream;
+use crate::{InputOrderMode, RecordBatchStream, SendableRecordBatchStream};
+
+/// Single aggregate stream for `InputOrderMode::Sorted` and
+/// `InputOrderMode::PartiallySorted`.
+///
+/// # Example
+///
+/// SELECT k, AVG(v) FROM t GROUP BY k;
+///
+/// If the input is ordered by `k`, and there are existing key partitioning on group
+/// by keys, the single mode aggregation with ordering optimization can be used:
+///
+/// ## Plan
+/// AggregateExec(stage=single, ordered)
+/// -- DataSourceExec(t)
+///
+/// ## Single Stage Behavior
+/// Input: raw rows
+/// Output: final results for all groups (for example, `AVG(x)`)
+///
+/// # Order-based Optimization
+///
+/// For the aggregation work, the hash aggregation implementation is reused.
+///
+/// After each input batch, check whether any groups can be emitted eagerly to
+/// improve memory efficiency. For example, if the last group key seen is
+/// `k = 100`, it is safe to emit all groups with keys less than 100 because the
+/// input is ordered.
+///
+/// # Memory Pressure and Spilling
+///
+/// ## Fully ordered case
+///
+/// If the input is ordered by every group key, for example:
+///
+/// - Input order: `a, b`
+/// - `GROUP BY`: `a, b`
+///
+/// Completed groups can be emitted as soon as the next group is observed. Thus,
+/// only the current group remains active after completed groups are emitted, and
+/// memory usage does not grow with the total number of groups.
+///
+/// If a memory reservation nevertheless fails, the stream returns the error
+/// directly, indicating an unexpected behavior.
+///
+/// ## Partially ordered case
+///
+/// If the input is ordered by only a subset of the group keys, for example:
+///
+/// - Input order: `a`
+/// - `GROUP BY`: `a, b`
+///
+/// If one `a` value contains many distinct `b` values, the table may accumulate
+/// enough groups to exceed the memory limit.
+///
+/// On reservation failure, the stream sorts the current intermediate states by
+/// the complete group key and spills them as one run. After the input ends, it
+/// spills any remaining states, performs a sort-preserving merge of all runs,
+/// and feeds the merged input into a fully ordered final aggregate stream.
+pub(crate) struct OrderedSingleAggregateStream {
+ schema: SchemaRef,
+ input: SendableRecordBatchStream,
+ reservation: MemoryReservation,
+ baseline_metrics: BaselineMetrics,
+ state: Option<OrderedSingleAggregateState>,
+}
+
+/// Spill configuration and accumulated runs for partially ordered single
+/// aggregation.
+///
+/// Each spill event drains all currently buffered groups, sorts their intermediate
+/// states by the full group key, and writes them to one spill file. All files are
+/// merged and replayed after the original input ends.
+struct OrderedSingleSpillContext {
+ /// Aggregate configuration used to construct the final replay stream.
+ final_agg: AggregateExec,
+ /// Task context
+ context: Arc<TaskContext>,
+ /// Original partition index
+ partition: usize,
+ /// Target batch size from configuration
+ batch_size: usize,
+ /// Full group-key ordering, such ordering with be kept in: a) individual spill
+ /// files, b) order after final merging and streaming aggregate
+ spill_expr: LexOrdering,
+ /// Spill I/O and metrics manager.
+ spill_manager: SpillManager,
+ /// Fully sorted spill runs waiting to be merged.
+ spills: Vec<SortedSpillFile>,
+}
+
+/// See comments at `poll_next()` for details.
+enum OrderedSingleAggregateState {
+ ReadingInput {
+ table: OrderedAggregateTable<SingleMarker>,
+ /// None if either
+ /// - Disk Manager doesn't enable temporary file creation
+ /// - The group keys are fully ordered, it's expected to use bounded memory
+ spill_context: Option<Box<OrderedSingleSpillContext>>,
+ },
+ Spilling {
+ table: OrderedAggregateTable<SingleMarker>,
+ spill_context: Box<OrderedSingleSpillContext>,
+ },
+ ProducingOutput {
+ table: OrderedAggregateTable<SingleMarker>,
+ },
+ PreparingMergeInput {
+ table: OrderedAggregateTable<SingleMarker>,
+ spill_context: Box<OrderedSingleSpillContext>,
+ },
+ MergingSpills {
+ stream: SendableRecordBatchStream,
+ },
+ Done,
+ /// Sentinel state to use when returning error from any other states, because:
+ /// - It explicitly releases state-owned resources immediately
+ /// - More defensive against accidentally resuming execution after error
+ Error,
+}
+
+type OrderedSingleAggregatePoll = Poll<Option<Result<RecordBatch>>>;
+type OrderedSingleAggregateStateTransition = ControlFlow<
+ (OrderedSingleAggregatePoll, OrderedSingleAggregateState),
+ OrderedSingleAggregateState,
+>;
+
+impl OrderedSingleSpillContext {
+ fn new(
+ agg: &AggregateExec,
+ context: &Arc<TaskContext>,
+ partition: usize,
+ batch_size: usize,
+ input_order_mode: &InputOrderMode,
+ spill_schema: &SchemaRef,
+ spill_metrics: SpillMetrics,
+ ) -> Result<Self> {
+ let group_schema = agg.group_by.group_schema(&agg.input().schema())?;
+ let output_ordering = agg.cache.output_ordering();
+ let InputOrderMode::PartiallySorted(order_indices) = input_order_mode else {
+ return internal_err!(
+ "Ordered single spill requires partially ordered input"
+ );
+ };
+ let spill_indices = order_indices.iter().copied().chain(
+ (0..group_schema.fields().len()).filter(|idx| !order_indices.contains(idx)),
+ );
+ let spill_sort_exprs = spill_indices.map(|idx| {
+ let field = group_schema.field(idx);
+ let output_expr = Column::new(field.name(), idx);
+ let sort_options = output_ordering
+ .and_then(|ordering| ordering.get_sort_options(&output_expr))
+ .unwrap_or_default();
+ PhysicalSortExpr::new(Arc::new(output_expr), sort_options)
+ });
+ let Some(spill_expr) = LexOrdering::new(spill_sort_exprs) else {
+ return internal_err!("Ordered single spill expression is empty");
+ };
+
+ let spill_manager = SpillManager::new(
+ context.runtime_env(),
+ spill_metrics,
+ Arc::clone(spill_schema),
+ )
+ .with_compression_type(context.session_config().spill_compression());
+
+ // Spilled rows contain group keys and intermediate states. Replay must
+ // merge those states and evaluate the final aggregate values.
+ let mut final_agg = agg.clone();
+ final_agg.mode = match agg.mode {
+ AggregateMode::Single => AggregateMode::Final,
+ AggregateMode::SinglePartitioned => AggregateMode::FinalPartitioned,
+ mode => {
+ return internal_err!(
+ "Ordered single aggregate spill cannot replay aggregate mode {mode:?}"
+ );
+ }
+ };
+ final_agg.group_by = Arc::new(agg.group_by.as_final());
+ final_agg.input_order_mode = InputOrderMode::Sorted;
+
+ Ok(Self {
+ final_agg,
+ context: Arc::clone(context),
+ partition,
+ batch_size,
+ spill_expr,
+ spill_manager,
+ spills: vec![],
+ })
+ }
+
+ fn has_spills(&self) -> bool {
+ !self.spills.is_empty()
+ }
+
+ /// Sorts and spills the aggregated groups. Memory reservation should be updated
+ /// by the caller.
+ ///
+ /// Individual spill files are ordered by the `group by` keys.
+ ///
+ /// See [`OrderedSingleAggregateStream`] for spilling details.
+ fn spill_table(
+ &mut self,
+ table: &mut OrderedAggregateTable<SingleMarker>,
+ ) -> Result<()> {
+ let Some(batch) = table.take_state_batch()? else {
+ return Ok(());
+ };
+
+ let sorted_iter =
+ IncrementalSortIterator::new(batch, self.spill_expr.clone(), self.batch_size);
+ let spill_file = self
+ .spill_manager
+ .spill_record_batch_iter_and_return_max_batch_memory(
+ sorted_iter,
+ "OrderedSingleAggregateSpill",
+ )?;
+
+ let Some((file, max_record_batch_memory)) = spill_file else {
+ return internal_err!("Ordered single aggregation produced an empty spill");
+ };
+
+ self.spills.push(SortedSpillFile {
+ file,
+ max_record_batch_memory,
+ });
+
+ Ok(())
+ }
+
+ /// Merges every sorted run and finalizes it through the fully ordered path.
+ fn into_replay_stream(
Review Comment:
This step is mostly similar to the spilling paths of other aggregation modes. We should extract the common logic later.
--
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@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] refactor(hash-aggr): migrate ordered single aggregation [datafusion]
Posted by "2010YOUY01 (via GitHub)" <gi...@apache.org>.
2010YOUY01 merged PR #24259:
URL: https://github.com/apache/datafusion/pull/24259
--
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@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] refactor(hash-aggr): migrate ordered single aggregation [datafusion]
Posted by "codecov-commenter (via GitHub)" <gi...@apache.org>.
codecov-commenter commented on PR #24259:
URL: https://github.com/apache/datafusion/pull/24259#issuecomment-5252268536
## [Codecov](https://app.codecov.io/gh/apache/datafusion/pull/24259?dropdown=coverage&src=pr&el=h1&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) Report
:x: Patch coverage is `86.86515%` with `75 lines` in your changes missing coverage. Please review.
:white_check_mark: Project coverage is 81.00%. Comparing base ([`634f0b3`](https://app.codecov.io/gh/apache/datafusion/commit/634f0b3004af274b0a3d8230710090447c53f352?dropdown=coverage&el=desc&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache)) to head ([`da5eeb6`](https://app.codecov.io/gh/apache/datafusion/commit/da5eeb66178d2f940ce4c7c0092e5a58a23fd02e?dropdown=coverage&el=desc&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache)).
:warning: Report is 42 commits behind head on main.
| [Files with missing lines](https://app.codecov.io/gh/apache/datafusion/pull/24259?dropdown=coverage&src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) | Patch % | Lines |
|---|---|---|
| [...sical-plan/src/aggregates/ordered\_single\_stream.rs](https://app.codecov.io/gh/apache/datafusion/pull/24259?src=pr&el=tree&filepath=datafusion%2Fphysical-plan%2Fsrc%2Faggregates%2Fordered_single_stream.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1wbGFuL3NyYy9hZ2dyZWdhdGVzL29yZGVyZWRfc2luZ2xlX3N0cmVhbS5ycw==) | 86.74% | [49 Missing and 10 partials :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/24259?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [datafusion/physical-plan/src/aggregates/mod.rs](https://app.codecov.io/gh/apache/datafusion/pull/24259?src=pr&el=tree&filepath=datafusion%2Fphysical-plan%2Fsrc%2Faggregates%2Fmod.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1wbGFuL3NyYy9hZ2dyZWdhdGVzL21vZC5ycw==) | 84.41% | [4 Missing and 8 partials :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/24259?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [.../aggregates/aggregate\_hash\_table/common\_ordered.rs](https://app.codecov.io/gh/apache/datafusion/pull/24259?src=pr&el=tree&filepath=datafusion%2Fphysical-plan%2Fsrc%2Faggregates%2Faggregate_hash_table%2Fcommon_ordered.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1wbGFuL3NyYy9hZ2dyZWdhdGVzL2FnZ3JlZ2F0ZV9oYXNoX3RhYmxlL2NvbW1vbl9vcmRlcmVkLnJz) | 60.00% | [0 Missing and 2 partials :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/24259?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
| [...gates/aggregate\_hash\_table/ordered\_single\_table.rs](https://app.codecov.io/gh/apache/datafusion/pull/24259?src=pr&el=tree&filepath=datafusion%2Fphysical-plan%2Fsrc%2Faggregates%2Faggregate_hash_table%2Fordered_single_table.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache#diff-ZGF0YWZ1c2lvbi9waHlzaWNhbC1wbGFuL3NyYy9hZ2dyZWdhdGVzL2FnZ3JlZ2F0ZV9oYXNoX3RhYmxlL29yZGVyZWRfc2luZ2xlX3RhYmxlLnJz) | 94.44% | [0 Missing and 2 partials :warning: ](https://app.codecov.io/gh/apache/datafusion/pull/24259?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache) |
<details><summary>Additional details and impacted files</summary>
```diff
@@ Coverage Diff @@
## main #24259 +/- ##
==========================================
- Coverage 81.06% 81.00% -0.06%
==========================================
Files 1107 1108 +1
Lines 382191 384599 +2408
Branches 382191 384599 +2408
==========================================
+ Hits 309805 311533 +1728
- Misses 54083 54709 +626
- Partials 18303 18357 +54
```
</details>
[:umbrella: View full report in Codecov by Harness](https://app.codecov.io/gh/apache/datafusion/pull/24259?dropdown=coverage&src=pr&el=continue&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache).
:loudspeaker: Have feedback on the report? [Share it here](https://about.codecov.io/codecov-pr-comment-feedback/?utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=apache).
<details><summary> :rocket: New features to boost your workflow: </summary>
- :snowflake: [Test Analytics](https://docs.codecov.com/docs/test-analytics): Detect flaky tests, report on failures, and find test suite problems.
- :package: [JS Bundle Analysis](https://docs.codecov.com/docs/javascript-bundle-analysis): Save yourself from yourself by tracking and limiting bundle sizes in JS merges.
</details>
--
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@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] refactor(hash-aggr): migrate ordered single aggregation [datafusion]
Posted by "2010YOUY01 (via GitHub)" <gi...@apache.org>.
2010YOUY01 commented on code in PR #24259:
URL: https://github.com/apache/datafusion/pull/24259#discussion_r3772026762
##########
datafusion/physical-plan/src/aggregates/ordered_single_stream.rs:
##########
@@ -0,0 +1,886 @@
+// Licensed to the Apache Software Foundation (ASF) under one
+// or more contributor license agreements. See the NOTICE file
+// distributed with this work for additional information
+// regarding copyright ownership. The ASF licenses this file
+// to you under the Apache License, Version 2.0 (the
+// "License"); you may not use this file except in compliance
+// with the License. You may obtain a copy of the License at
+//
+// http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing,
+// software distributed under the License is distributed on an
+// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, either express or implied. See the License for the
+// specific language governing permissions and limitations
+// under the License.
+
+//! Single-stage aggregate stream for ordered raw input.
+
+use std::ops::ControlFlow;
+use std::sync::Arc;
+use std::task::{Context, Poll};
+
+use arrow::datatypes::SchemaRef;
+use arrow::record_batch::RecordBatch;
+use datafusion_common::{DataFusionError, Result, internal_datafusion_err, internal_err};
+use datafusion_execution::TaskContext;
+use datafusion_execution::memory_pool::{MemoryConsumer, MemoryReservation};
+use datafusion_physical_expr::PhysicalSortExpr;
+use datafusion_physical_expr::expressions::Column;
+use datafusion_physical_expr_common::sort_expr::LexOrdering;
+use futures::stream::{Stream, StreamExt};
+
+use super::aggregate_hash_table::{OrderedAggregateTable, SingleMarker};
+use super::group_values::GroupByMetrics;
+use super::ordered_final_stream::OrderedFinalAggregateStream;
+use super::{AggregateExec, create_schema};
+use crate::aggregates::AggregateMode;
+use crate::metrics::{BaselineMetrics, RecordOutput, SpillMetrics};
+use crate::sorts::IncrementalSortIterator;
+use crate::sorts::streaming_merge::{SortedSpillFile, StreamingMergeBuilder};
+use crate::spill::spill_manager::SpillManager;
+use crate::stream::EmptyRecordBatchStream;
+use crate::{InputOrderMode, RecordBatchStream, SendableRecordBatchStream};
+
+/// Single aggregate stream for `InputOrderMode::Sorted` and
+/// `InputOrderMode::PartiallySorted`.
+///
+/// # Example
+///
+/// SELECT k, AVG(v) FROM t GROUP BY k;
+///
+/// If the input is ordered by `k`, and there are existing key partitioning on group
+/// by keys, the single mode aggregation with ordering optimization can be used:
+///
+/// ## Plan
+/// AggregateExec(stage=single, ordered)
+/// -- DataSourceExec(t)
+///
+/// ## Single Stage Behavior
+/// Input: raw rows
+/// Output: final results for all groups (for example, `AVG(x)`)
+///
+/// # Order-based Optimization
+///
+/// For the aggregation work, the hash aggregation implementation is reused.
+///
+/// After each input batch, check whether any groups can be emitted eagerly to
+/// improve memory efficiency. For example, if the last group key seen is
+/// `k = 100`, it is safe to emit all groups with keys less than 100 because the
+/// input is ordered.
+///
+/// # Memory Pressure and Spilling
+///
+/// ## Fully ordered case
+///
+/// If the input is ordered by every group key, for example:
+///
+/// - Input order: `a, b`
+/// - `GROUP BY`: `a, b`
+///
+/// Completed groups can be emitted as soon as the next group is observed. Thus,
+/// only the current group remains active after completed groups are emitted, and
+/// memory usage does not grow with the total number of groups.
+///
+/// If a memory reservation nevertheless fails, the stream returns the error
+/// directly, indicating an unexpected behavior.
+///
+/// ## Partially ordered case
+///
+/// If the input is ordered by only a subset of the group keys, for example:
+///
+/// - Input order: `a`
+/// - `GROUP BY`: `a, b`
+///
+/// If one `a` value contains many distinct `b` values, the table may accumulate
+/// enough groups to exceed the memory limit.
+///
+/// On reservation failure, the stream sorts the current intermediate states by
+/// the complete group key and spills them as one run. After the input ends, it
+/// spills any remaining states, performs a sort-preserving merge of all runs,
+/// and feeds the merged input into a fully ordered final aggregate stream.
+pub(crate) struct OrderedSingleAggregateStream {
+ schema: SchemaRef,
+ input: SendableRecordBatchStream,
+ reservation: MemoryReservation,
+ baseline_metrics: BaselineMetrics,
+ state: Option<OrderedSingleAggregateState>,
+}
+
+/// Spill configuration and accumulated runs for partially ordered single
+/// aggregation.
+///
+/// Each spill event drains all currently buffered groups, sorts their intermediate
+/// states by the full group key, and writes them to one spill file. All files are
+/// merged and replayed after the original input ends.
+struct OrderedSingleSpillContext {
+ /// Aggregate configuration used to construct the final replay stream.
+ final_agg: AggregateExec,
+ /// Task context
+ context: Arc<TaskContext>,
+ /// Original partition index
+ partition: usize,
+ /// Target batch size from configuration
+ batch_size: usize,
+ /// Full group-key ordering, such ordering with be kept in: a) individual spill
+ /// files, b) order after final merging and streaming aggregate
+ spill_expr: LexOrdering,
+ /// Spill I/O and metrics manager.
+ spill_manager: SpillManager,
+ /// Fully sorted spill runs waiting to be merged.
+ spills: Vec<SortedSpillFile>,
+}
+
+/// See comments at `poll_next()` for details.
+enum OrderedSingleAggregateState {
+ ReadingInput {
+ table: OrderedAggregateTable<SingleMarker>,
+ /// None if either
+ /// - Disk Manager doesn't enable temporary file creation
+ /// - The group keys are fully ordered, it's expected to use bounded memory
+ spill_context: Option<Box<OrderedSingleSpillContext>>,
+ },
+ Spilling {
+ table: OrderedAggregateTable<SingleMarker>,
+ spill_context: Box<OrderedSingleSpillContext>,
+ },
+ ProducingOutput {
+ table: OrderedAggregateTable<SingleMarker>,
+ },
+ PreparingMergeInput {
+ table: OrderedAggregateTable<SingleMarker>,
+ spill_context: Box<OrderedSingleSpillContext>,
+ },
+ MergingSpills {
+ stream: SendableRecordBatchStream,
+ },
+ Done,
+ /// Sentinel state to use when returning error from any other states, because:
+ /// - It explicitly releases state-owned resources immediately
+ /// - More defensive against accidentally resuming execution after error
+ Error,
+}
+
+type OrderedSingleAggregatePoll = Poll<Option<Result<RecordBatch>>>;
+type OrderedSingleAggregateStateTransition = ControlFlow<
+ (OrderedSingleAggregatePoll, OrderedSingleAggregateState),
+ OrderedSingleAggregateState,
+>;
+
+impl OrderedSingleSpillContext {
+ fn new(
+ agg: &AggregateExec,
+ context: &Arc<TaskContext>,
+ partition: usize,
+ batch_size: usize,
+ input_order_mode: &InputOrderMode,
+ spill_schema: &SchemaRef,
+ spill_metrics: SpillMetrics,
+ ) -> Result<Self> {
+ let group_schema = agg.group_by.group_schema(&agg.input().schema())?;
+ let output_ordering = agg.cache.output_ordering();
+ let InputOrderMode::PartiallySorted(order_indices) = input_order_mode else {
+ return internal_err!(
+ "Ordered single spill requires partially ordered input"
+ );
+ };
+ let spill_indices = order_indices.iter().copied().chain(
+ (0..group_schema.fields().len()).filter(|idx| !order_indices.contains(idx)),
+ );
+ let spill_sort_exprs = spill_indices.map(|idx| {
+ let field = group_schema.field(idx);
+ let output_expr = Column::new(field.name(), idx);
+ let sort_options = output_ordering
+ .and_then(|ordering| ordering.get_sort_options(&output_expr))
+ .unwrap_or_default();
+ PhysicalSortExpr::new(Arc::new(output_expr), sort_options)
+ });
+ let Some(spill_expr) = LexOrdering::new(spill_sort_exprs) else {
+ return internal_err!("Ordered single spill expression is empty");
+ };
+
+ let spill_manager = SpillManager::new(
+ context.runtime_env(),
+ spill_metrics,
+ Arc::clone(spill_schema),
+ )
+ .with_compression_type(context.session_config().spill_compression());
+
+ // Spilled rows contain group keys and intermediate states. Replay must
+ // merge those states and evaluate the final aggregate values.
+ let mut final_agg = agg.clone();
+ final_agg.mode = match agg.mode {
+ AggregateMode::Single => AggregateMode::Final,
+ AggregateMode::SinglePartitioned => AggregateMode::FinalPartitioned,
+ mode => {
+ return internal_err!(
+ "Ordered single aggregate spill cannot replay aggregate mode {mode:?}"
+ );
+ }
+ };
+ final_agg.group_by = Arc::new(agg.group_by.as_final());
+ final_agg.input_order_mode = InputOrderMode::Sorted;
+
+ Ok(Self {
+ final_agg,
+ context: Arc::clone(context),
+ partition,
+ batch_size,
+ spill_expr,
+ spill_manager,
+ spills: vec![],
+ })
+ }
+
+ fn has_spills(&self) -> bool {
+ !self.spills.is_empty()
+ }
+
+ /// Sorts and spills the aggregated groups. Memory reservation should be updated
+ /// by the caller.
+ ///
+ /// Individual spill files are ordered by the `group by` keys.
+ ///
+ /// See [`OrderedSingleAggregateStream`] for spilling details.
+ fn spill_table(
+ &mut self,
+ table: &mut OrderedAggregateTable<SingleMarker>,
+ ) -> Result<()> {
+ let Some(batch) = table.take_state_batch()? else {
+ return Ok(());
+ };
+
+ let sorted_iter =
+ IncrementalSortIterator::new(batch, self.spill_expr.clone(), self.batch_size);
+ let spill_file = self
+ .spill_manager
+ .spill_record_batch_iter_and_return_max_batch_memory(
+ sorted_iter,
+ "OrderedSingleAggregateSpill",
+ )?;
+
+ let Some((file, max_record_batch_memory)) = spill_file else {
+ return internal_err!("Ordered single aggregation produced an empty spill");
+ };
+
+ self.spills.push(SortedSpillFile {
+ file,
+ max_record_batch_memory,
+ });
+
+ Ok(())
+ }
+
+ /// Merges every sorted run and finalizes it through the fully ordered path.
+ fn into_replay_stream(
Review Comment:
https://github.com/apache/datafusion/issues/24317
--
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@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org
Re: [PR] refactor(hash-aggr): migrate ordered single aggregation [datafusion]
Posted by "2010YOUY01 (via GitHub)" <gi...@apache.org>.
2010YOUY01 commented on PR #24259:
URL: https://github.com/apache/datafusion/pull/24259#issuecomment-5275698545
Thanks for the review @alamb
--
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@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscribe@datafusion.apache.org
For additional commands, e-mail: github-help@datafusion.apache.org