Skip to content

fix: median returns Float64 for integer inputs to avoid truncation #21988

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 1 commit into from
May 6, 2026
Merged

fix: median returns Float64 for integer inputs to avoid truncation #21988

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
2 changes: 1 addition & 1 deletion datafusion/core/tests/dataframe/describe.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -44,7 +44,7 @@ async fn describe() -> Result<()> {
| std | 2107.472815166704 | null | 2.8724780750809518 | 2.8724780750809518 | 2.8724780750809518 | 28.724780750809533 | 3.1597258182544645 | 29.012028558317645 | null | null | null | 0.5000342500942125 | 3.44808750051728 |
| min | 0.0 | null | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 01/01/09 | 0 | 2008-12-31T23:00:00 | 2009.0 | 1.0 |
| max | 7299.0 | null | 9.0 | 9.0 | 9.0 | 90.0 | 9.899999618530273 | 90.89999999999999 | 12/31/10 | 9 | 2010-12-31T04:09:13.860 | 2010.0 | 12.0 |
| median | 3649.0 | null | 4.0 | 4.0 | 4.0 | 45.0 | 4.949999809265137 | 45.45 | null | null | null | 2009.0 | 7.0 |
| median | 3649.5 | null | 4.5 | 4.5 | 4.5 | 45.0 | 4.949999809265137 | 45.45 | null | null | null | 2009.5 | 7.0 |
+------------+-------------------+----------+--------------------+--------------------+--------------------+--------------------+--------------------+--------------------+-----------------+------------+-------------------------+--------------------+-------------------+
");
Ok(())
Expand Down
40 changes: 20 additions & 20 deletions datafusion/core/tests/dataframe/mod.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1204,26 +1204,26 @@ async fn window_using_aggregates() -> Result<()> {
| first_value | last_val | approx_distinct | approx_median | median | max | min | c2 | c3 |
+-------------+----------+-----------------+---------------+--------+-----+------+----+------+
| | | | | | | | 1 | -85 |
| -85 | -101 | 14 | -12.0 | -12 | 83 | -101 | 4 | -54 |
| -85 | -101 | 17 | -25.0 | -25 | 83 | -101 | 5 | -31 |
| -85 | -12 | 10 | -32.75 | -34 | 83 | -85 | 3 | 13 |
| -85 | -25 | 3 | -56.0 | -56 | -25 | -85 | 1 | -5 |
| -85 | -31 | 18 | -29.75 | -28 | 83 | -101 | 5 | 36 |
| -85 | -38 | 16 | -25.0 | -25 | 83 | -101 | 4 | 65 |
| -85 | -43 | 7 | -43.0 | -43 | 83 | -85 | 2 | 45 |
| -85 | -48 | 6 | -35.75 | -36 | 83 | -85 | 2 | -43 |
| -85 | -5 | 4 | -37.75 | -40 | -5 | -85 | 1 | 83 |
| -85 | -54 | 15 | -17.0 | -18 | 83 | -101 | 4 | -38 |
| -85 | -56 | 2 | -70.5 | -70 | -56 | -85 | 1 | -25 |
| -85 | -72 | 9 | -43.0 | -43 | 83 | -85 | 3 | -12 |
| -85 | -85 | 1 | -85.0 | -85 | -85 | -85 | 1 | -56 |
| -85 | 13 | 11 | -17.0 | -18 | 83 | -85 | 3 | 14 |
| -85 | 13 | 11 | -25.0 | -25 | 83 | -85 | 3 | 13 |
| -85 | 14 | 12 | -12.0 | -12 | 83 | -85 | 3 | 17 |
| -85 | 17 | 13 | -11.25 | -8 | 83 | -85 | 4 | -101 |
| -85 | 45 | 8 | -34.5 | -34 | 83 | -85 | 3 | -72 |
| -85 | 65 | 17 | -17.0 | -18 | 83 | -101 | 5 | -101 |
| -85 | 83 | 5 | -25.0 | -25 | 83 | -85 | 2 | -48 |
| -85 | -101 | 14 | -12.0 | -12.0 | 83 | -101 | 4 | -54 |
| -85 | -101 | 17 | -25.0 | -25.0 | 83 | -101 | 5 | -31 |
| -85 | -12 | 10 | -32.75 | -34.0 | 83 | -85 | 3 | 13 |
| -85 | -25 | 3 | -56.0 | -56.0 | -25 | -85 | 1 | -5 |
| -85 | -31 | 18 | -29.75 | -28.0 | 83 | -101 | 5 | 36 |
| -85 | -38 | 16 | -25.0 | -25.0 | 83 | -101 | 4 | 65 |
| -85 | -43 | 7 | -43.0 | -43.0 | 83 | -85 | 2 | 45 |
| -85 | -48 | 6 | -35.75 | -36.5 | 83 | -85 | 2 | -43 |
| -85 | -5 | 4 | -37.75 | -40.5 | -5 | -85 | 1 | 83 |
| -85 | -54 | 15 | -17.0 | -18.5 | 83 | -101 | 4 | -38 |
| -85 | -56 | 2 | -70.5 | -70.5 | -56 | -85 | 1 | -25 |
| -85 | -72 | 9 | -43.0 | -43.0 | 83 | -85 | 3 | -12 |
| -85 | -85 | 1 | -85.0 | -85.0 | -85 | -85 | 1 | -56 |
| -85 | 13 | 11 | -17.0 | -18.5 | 83 | -85 | 3 | 14 |
| -85 | 13 | 11 | -25.0 | -25.0 | 83 | -85 | 3 | 13 |
| -85 | 14 | 12 | -12.0 | -12.0 | 83 | -85 | 3 | 17 |
| -85 | 17 | 13 | -11.25 | -8.5 | 83 | -85 | 4 | -101 |
| -85 | 45 | 8 | -34.5 | -34.0 | 83 | -85 | 3 | -72 |
| -85 | 65 | 17 | -17.0 | -18.5 | 83 | -101 | 5 | -101 |
| -85 | 83 | 5 | -25.0 | -25.0 | 83 | -85 | 2 | -48 |
+-------------+----------+-----------------+---------------+--------+-----+------+----+------+
"
);
Expand Down
28 changes: 14 additions & 14 deletions datafusion/core/tests/sql/aggregates/dict_nulls.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -91,10 +91,10 @@ async fn test_aggregates_null_handling_comprehensive() -> Result<()> {
+----------------+--------------+
| dict_null_vals | median_value |
+----------------+--------------+
| | 3 |
| group_x | 1 |
| group_y | 5 |
| group_z | 7 |
| | 3.0 |
| group_x | 1.0 |
| group_y | 5.0 |
| group_z | 7.0 |
+----------------+--------------+
");

Expand Down Expand Up @@ -437,16 +437,16 @@ async fn test_median_distinct_with_fuzz_table_dict_nulls() -> Result<()> {
assert_snapshot!(
batches_to_string(&results),
@r"
+--------+---------------------+------+------+------+--------+--------+
| u8_low | dictionary_utf8_low | col1 | col2 | col3 | col4 | col5 |
+--------+---------------------+------+------+------+--------+--------+
| 50 | | | 30 | | 987.65 | 400000 |
| 50 | group_three | 5000 | 50 | 5000 | 555.55 | 500000 |
| 75 | | 4000 | | 4000 | | 450000 |
| 100 | group_one | 1100 | 11 | 1000 | 123.45 | 110000 |
| 100 | group_two | 1500 | 15 | 1500 | 111.11 | 150000 |
| 200 | | 2500 | 22 | 2500 | 506.11 | 250000 |
+--------+---------------------+------+------+------+--------+--------+
+--------+---------------------+--------+------+--------+--------+----------+
| u8_low | dictionary_utf8_low | col1 | col2 | col3 | col4 | col5 |
+--------+---------------------+--------+------+--------+--------+----------+
| 50 | | | 30.0 | | 987.65 | 400000.0 |
| 50 | group_three | 5000.0 | 50.0 | 5000.0 | 555.55 | 500000.0 |
| 75 | | 4000.0 | | 4000.0 | | 450000.0 |
| 100 | group_one | 1100.0 | 11.0 | 1000.0 | 123.45 | 110000.0 |
| 100 | group_two | 1500.0 | 15.0 | 1500.0 | 111.11 | 150000.0 |
| 200 | | 2500.0 | 22.5 | 2500.0 | 506.11 | 250000.0 |
+--------+---------------------+--------+------+--------+--------+----------+
"
);

Expand Down
25 changes: 22 additions & 3 deletions datafusion/functions-aggregate/src/median.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -39,14 +39,15 @@ use arrow::datatypes::{
ArrowNativeType, ArrowPrimitiveType, Decimal32Type, Decimal64Type, FieldRef,
};

use datafusion_common::types::{NativeType, logical_float64};
use datafusion_common::{
DataFusionError, Result, ScalarValue, assert_eq_or_internal_err,
internal_datafusion_err,
};
use datafusion_expr::function::StateFieldsArgs;
use datafusion_expr::{
Accumulator, AggregateUDFImpl, Documentation, Signature, Volatility,
function::AccumulatorArgs, utils::format_state_name,
Accumulator, AggregateUDFImpl, Coercion, Documentation, Signature, TypeSignature,
TypeSignatureClass, Volatility, function::AccumulatorArgs, utils::format_state_name,
};
use datafusion_expr::{EmitTo, GroupsAccumulator};
use datafusion_functions_aggregate_common::aggregate::groups_accumulator::accumulate::accumulate;
Expand Down Expand Up @@ -99,7 +100,25 @@ impl Default for Median {
impl Median {
pub fn new() -> Self {
Self {
signature: Signature::numeric(1, Volatility::Immutable),
// Integer inputs are coerced to Float64 so the average of the two
// middle values is not truncated. This matches DuckDB / PostgreSQL / Spark.
// Float and Decimal inputs preserve their type.
signature: Signature::one_of(
vec![
TypeSignature::Coercible(vec![Coercion::new_exact(
TypeSignatureClass::Decimal,
)]),
TypeSignature::Coercible(vec![Coercion::new_exact(
TypeSignatureClass::Float,
)]),
TypeSignature::Coercible(vec![Coercion::new_implicit(
TypeSignatureClass::Native(logical_float64()),
vec![TypeSignatureClass::Integer],
NativeType::Float64,
)]),
],
Volatility::Immutable,
),
}
}
}
Expand Down
Loading
Loading