Skip to content

Implement retract_batch for AvgAccumulator #4846

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.

Sign up for GitHub

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

Merged
alamb merged 2 commits into from
Jan 10, 2023
Merged

Implement retract_batch for AvgAccumulator #4846

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
eb093a8
Implement retract_batch for AvgAccumulator,
jonmmease Jan 7, 2023
830a6d0
fmt
jonmmease Jan 7, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
38 changes: 20 additions & 18 deletions datafusion/core/tests/sql/window.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -503,21 +503,22 @@ async fn window_frame_rows_preceding() -> Result<()> {
register_aggregate_csv(&ctx).await?;
let sql = "SELECT \
SUM(c4) OVER(ORDER BY c4 ROWS BETWEEN 1 PRECEDING AND 1 FOLLOWING),\
AVG(c4) OVER(ORDER BY c4 ROWS BETWEEN 1 PRECEDING AND 1 FOLLOWING),\
COUNT(*) OVER(ORDER BY c4 ROWS BETWEEN 1 PRECEDING AND 1 FOLLOWING)\
FROM aggregate_test_100 \
ORDER BY c9 \
LIMIT 5";
let actual = execute_to_batches(&ctx, sql).await;
let expected = vec![
"+----------------------------+-----------------+",
"| SUM(aggregate_test_100.c4) | COUNT(UInt8(1)) |",
"+----------------------------+-----------------+",
"| -48302 | 3 |",
"| 11243 | 3 |",
"| -51311 | 3 |",
"| -2391 | 3 |",
"| 46756 | 3 |",
"+----------------------------+-----------------+",
"+----------------------------+----------------------------+-----------------+",
"| SUM(aggregate_test_100.c4) | AVG(aggregate_test_100.c4) | COUNT(UInt8(1)) |",
"+----------------------------+----------------------------+-----------------+",
"| -48302 | -16100.666666666666 | 3 |",
"| 11243 | 3747.6666666666665 | 3 |",
"| -51311 | -17103.666666666668 | 3 |",
"| -2391 | -797 | 3 |",
"| 46756 | 15585.333333333334 | 3 |",
"+----------------------------+----------------------------+-----------------+",
];
assert_batches_eq!(expected, &actual);
Ok(())
Expand All @@ -529,21 +530,22 @@ async fn window_frame_rows_preceding_with_partition_unique_order_by() -> Result<
register_aggregate_csv(&ctx).await?;
let sql = "SELECT \
SUM(c4) OVER(PARTITION BY c1 ORDER BY c9 ROWS BETWEEN 1 PRECEDING AND 1 FOLLOWING),\
AVG(c4) OVER(PARTITION BY c1 ORDER BY c9 ROWS BETWEEN 1 PRECEDING AND 1 FOLLOWING),\
COUNT(*) OVER(PARTITION BY c2 ORDER BY c9 ROWS BETWEEN 1 PRECEDING AND 1 FOLLOWING)\
FROM aggregate_test_100 \
ORDER BY c9 \
LIMIT 5";
let actual = execute_to_batches(&ctx, sql).await;
let expected = vec![
"+----------------------------+-----------------+",
"| SUM(aggregate_test_100.c4) | COUNT(UInt8(1)) |",
"+----------------------------+-----------------+",
"| -38611 | 2 |",
"| 17547 | 2 |",
"| -1301 | 2 |",
"| 26638 | 3 |",
"| 26861 | 3 |",
"+----------------------------+-----------------+",
"+----------------------------+----------------------------+-----------------+",
"| SUM(aggregate_test_100.c4) | AVG(aggregate_test_100.c4) | COUNT(UInt8(1)) |",
"+----------------------------+----------------------------+-----------------+",
"| -38611 | -19305.5 | 2 |",
"| 17547 | 8773.5 | 2 |",
"| -1301 | -650.5 | 2 |",
"| 26638 | 13319 | 3 |",
"| 26861 | 8953.666666666666 | 3 |",
"+----------------------------+----------------------------+-----------------+",
];
assert_batches_eq!(expected, &actual);
Ok(())
Expand Down
13 changes: 13 additions & 0 deletions datafusion/physical-expr/src/aggregate/average.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,6 +25,7 @@ use crate::aggregate::row_accumulator::{
is_row_accumulator_support_dtype, RowAccumulator,
};
use crate::aggregate::sum;
use crate::aggregate::sum::sum_batch;
use crate::expressions::format_state_name;
use crate::{AggregateExpr, PhysicalExpr};
use arrow::compute;
Expand Down Expand Up @@ -119,6 +120,10 @@ impl AggregateExpr for Avg {
self.data_type.clone(),
)))
}

fn create_sliding_accumulator(&self) -> Result<Box<dyn Accumulator>> {
Ok(Box::new(AvgAccumulator::try_new(&self.data_type)?))
}
}

/// An accumulator to compute the average
Expand Down Expand Up @@ -154,6 +159,14 @@ impl Accumulator for AvgAccumulator {
Ok(())
}

fn retract_batch(&mut self, values: &[ArrayRef]) -> Result<()> {
let values = &values[0];
self.count -= (values.len() - values.data().null_count()) as u64;
let delta = sum_batch(values, &self.sum.get_datatype())?;
self.sum = self.sum.sub(&delta)?;
Comment on lines +165 to +166
Copy link
Contributor

@ozankabak ozankabak Jan 7, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I like giving a name to the change (i.e. delta). I suggest doing the same in update_batch.

Ok(())
}

fn merge_batch(&mut self, states: &[ArrayRef]) -> Result<()> {
let counts = downcast_value!(states[0], UInt64Array);
// counts are summed
Expand Down