This is an automated email from the ASF dual-hosted git repository.
github-bot pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/datafusion.git
The following commit(s) were added to refs/heads/main by this push:
new 3b390d782e Refactor avg & sum signatures away from user defined
(#18769)
3b390d782e is described below
commit 3b390d782ea390f3dd080aecf3ea6083b6406381
Author: Jeffrey Vo <[email protected]>
AuthorDate: Thu Nov 20 13:12:46 2025 +1100
Refactor avg & sum signatures away from user defined (#18769)
## 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 #12725
## 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.
-->
Prefer to avoid user_defined for consistency in function definitions.
## 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.
-->
Refactor signature of avg & sum away from user_defined.
## 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.
## 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.
-->
---
datafusion/functions-aggregate/src/average.rs | 55 ++++++-------
datafusion/functions-aggregate/src/sum.rs | 110 ++++++++++++-------------
datafusion/spark/src/function/aggregate/avg.rs | 35 +++-----
3 files changed, 91 insertions(+), 109 deletions(-)
diff --git a/datafusion/functions-aggregate/src/average.rs
b/datafusion/functions-aggregate/src/average.rs
index bec1734e2e..f4b3e598c3 100644
--- a/datafusion/functions-aggregate/src/average.rs
+++ b/datafusion/functions-aggregate/src/average.rs
@@ -31,18 +31,15 @@ use arrow::datatypes::{
DECIMAL256_MAX_SCALE, DECIMAL32_MAX_PRECISION, DECIMAL32_MAX_SCALE,
DECIMAL64_MAX_PRECISION, DECIMAL64_MAX_SCALE,
};
-use datafusion_common::plan_err;
-use datafusion_common::{
- exec_err, not_impl_err, utils::take_function_args, Result, ScalarValue,
-};
+use datafusion_common::types::{logical_float64, NativeType};
+use datafusion_common::{exec_err, not_impl_err, Result, ScalarValue};
use datafusion_expr::function::{AccumulatorArgs, StateFieldsArgs};
use datafusion_expr::utils::format_state_name;
-use datafusion_expr::Volatility::Immutable;
use datafusion_expr::{
- Accumulator, AggregateUDFImpl, Documentation, EmitTo, Expr,
GroupsAccumulator,
- ReversedUDAF, Signature,
+ Accumulator, AggregateUDFImpl, Coercion, Documentation, EmitTo, Expr,
+ GroupsAccumulator, ReversedUDAF, Signature, TypeSignature,
TypeSignatureClass,
+ Volatility,
};
-
use datafusion_functions_aggregate_common::aggregate::avg_distinct::{
DecimalDistinctAvgAccumulator, Float64DistinctAvgAccumulator,
};
@@ -50,7 +47,6 @@ use
datafusion_functions_aggregate_common::aggregate::groups_accumulator::accumu
use
datafusion_functions_aggregate_common::aggregate::groups_accumulator::nulls::{
filtered_null_mask, set_nulls,
};
-
use datafusion_functions_aggregate_common::utils::DecimalAverager;
use datafusion_macros::user_doc;
use log::debug;
@@ -101,7 +97,24 @@ pub struct Avg {
impl Avg {
pub fn new() -> Self {
Self {
- signature: Signature::user_defined(Immutable),
+ // Supported types smallint, int, bigint, real, double precision,
decimal, or interval
+ // Refer to
https://www.postgresql.org/docs/8.2/functions-aggregate.html doc
+ signature: Signature::one_of(
+ vec![
+ TypeSignature::Coercible(vec![Coercion::new_exact(
+ TypeSignatureClass::Decimal,
+ )]),
+ TypeSignature::Coercible(vec![Coercion::new_exact(
+ TypeSignatureClass::Duration,
+ )]),
+ TypeSignature::Coercible(vec![Coercion::new_implicit(
+ TypeSignatureClass::Native(logical_float64()),
+ vec![TypeSignatureClass::Integer,
TypeSignatureClass::Float],
+ NativeType::Float64,
+ )]),
+ ],
+ Volatility::Immutable,
+ ),
aliases: vec![String::from("mean")],
}
}
@@ -126,28 +139,6 @@ impl AggregateUDFImpl for Avg {
&self.signature
}
- fn coerce_types(&self, arg_types: &[DataType]) -> Result<Vec<DataType>> {
- let [args] = take_function_args(self.name(), arg_types)?;
-
- // Supported types smallint, int, bigint, real, double precision,
decimal, or interval
- // Refer to
https://www.postgresql.org/docs/8.2/functions-aggregate.html doc
- fn coerced_type(data_type: &DataType) -> Result<DataType> {
- match &data_type {
- DataType::Decimal32(p, s) => Ok(DataType::Decimal32(*p, *s)),
- DataType::Decimal64(p, s) => Ok(DataType::Decimal64(*p, *s)),
- DataType::Decimal128(p, s) => Ok(DataType::Decimal128(*p, *s)),
- DataType::Decimal256(p, s) => Ok(DataType::Decimal256(*p, *s)),
- d if d.is_numeric() => Ok(DataType::Float64),
- DataType::Duration(time_unit) =>
Ok(DataType::Duration(*time_unit)),
- DataType::Dictionary(_, v) => coerced_type(v.as_ref()),
- _ => {
- plan_err!("Avg does not support inputs of type
{data_type}.")
- }
- }
- }
- Ok(vec![coerced_type(args)?])
- }
-
fn return_type(&self, arg_types: &[DataType]) -> Result<DataType> {
match &arg_types[0] {
DataType::Decimal32(precision, scale) => {
diff --git a/datafusion/functions-aggregate/src/sum.rs
b/datafusion/functions-aggregate/src/sum.rs
index 958553d78c..d726b9fad1 100644
--- a/datafusion/functions-aggregate/src/sum.rs
+++ b/datafusion/functions-aggregate/src/sum.rs
@@ -18,35 +18,31 @@
//! Defines `SUM` and `SUM DISTINCT` aggregate accumulators
use ahash::RandomState;
-use arrow::datatypes::DECIMAL32_MAX_PRECISION;
-use arrow::datatypes::DECIMAL64_MAX_PRECISION;
-use datafusion_expr::utils::AggregateOrderSensitivity;
-use datafusion_expr::Expr;
-use std::any::Any;
-use std::mem::size_of_val;
-
-use arrow::array::Array;
-use arrow::array::ArrowNativeTypeOp;
-use arrow::array::{ArrowNumericType, AsArray};
-use arrow::datatypes::{ArrowNativeType, FieldRef};
+use arrow::array::{Array, ArrayRef, ArrowNativeTypeOp, ArrowNumericType,
AsArray};
+use arrow::datatypes::Field;
use arrow::datatypes::{
- DataType, Decimal128Type, Decimal256Type, Decimal32Type, Decimal64Type,
Float64Type,
- Int64Type, UInt64Type, DECIMAL128_MAX_PRECISION, DECIMAL256_MAX_PRECISION,
+ ArrowNativeType, DataType, Decimal128Type, Decimal256Type, Decimal32Type,
+ Decimal64Type, FieldRef, Float64Type, Int64Type, UInt64Type,
+ DECIMAL128_MAX_PRECISION, DECIMAL256_MAX_PRECISION,
DECIMAL32_MAX_PRECISION,
+ DECIMAL64_MAX_PRECISION,
};
-use arrow::{array::ArrayRef, datatypes::Field};
-use datafusion_common::{
- exec_err, not_impl_err, utils::take_function_args, HashMap, Result,
ScalarValue,
+use datafusion_common::types::{
+ logical_float64, logical_int16, logical_int32, logical_int64, logical_int8,
+ logical_uint16, logical_uint32, logical_uint64, logical_uint8, NativeType,
};
-use datafusion_expr::function::AccumulatorArgs;
-use datafusion_expr::function::StateFieldsArgs;
-use datafusion_expr::utils::format_state_name;
+use datafusion_common::{exec_err, not_impl_err, HashMap, Result, ScalarValue};
+use datafusion_expr::function::{AccumulatorArgs, StateFieldsArgs};
+use datafusion_expr::utils::{format_state_name, AggregateOrderSensitivity};
use datafusion_expr::{
- Accumulator, AggregateUDFImpl, Documentation, GroupsAccumulator,
ReversedUDAF,
- SetMonotonicity, Signature, Volatility,
+ Accumulator, AggregateUDFImpl, Coercion, Documentation, Expr,
GroupsAccumulator,
+ ReversedUDAF, SetMonotonicity, Signature, TypeSignature,
TypeSignatureClass,
+ Volatility,
};
use
datafusion_functions_aggregate_common::aggregate::groups_accumulator::prim_op::PrimitiveGroupsAccumulator;
use
datafusion_functions_aggregate_common::aggregate::sum_distinct::DistinctSumAccumulator;
use datafusion_macros::user_doc;
+use std::any::Any;
+use std::mem::size_of_val;
make_udaf_expr_and_func!(
Sum,
@@ -130,7 +126,42 @@ pub struct Sum {
impl Sum {
pub fn new() -> Self {
Self {
- signature: Signature::user_defined(Volatility::Immutable),
+ // Refer to
https://www.postgresql.org/docs/8.2/functions-aggregate.html doc
+ // smallint, int, bigint, real, double precision, decimal, or
interval.
+ signature: Signature::one_of(
+ vec![
+ TypeSignature::Coercible(vec![Coercion::new_exact(
+ TypeSignatureClass::Decimal,
+ )]),
+ // Unsigned to u64
+ TypeSignature::Coercible(vec![Coercion::new_implicit(
+ TypeSignatureClass::Native(logical_uint64()),
+ vec![
+ TypeSignatureClass::Native(logical_uint8()),
+ TypeSignatureClass::Native(logical_uint16()),
+ TypeSignatureClass::Native(logical_uint32()),
+ ],
+ NativeType::UInt64,
+ )]),
+ // Signed to i64
+ TypeSignature::Coercible(vec![Coercion::new_implicit(
+ TypeSignatureClass::Native(logical_int64()),
+ vec![
+ TypeSignatureClass::Native(logical_int8()),
+ TypeSignatureClass::Native(logical_int16()),
+ TypeSignatureClass::Native(logical_int32()),
+ ],
+ NativeType::Int64,
+ )]),
+ // Floats to f64
+ TypeSignature::Coercible(vec![Coercion::new_implicit(
+ TypeSignatureClass::Native(logical_float64()),
+ vec![TypeSignatureClass::Float],
+ NativeType::Float64,
+ )]),
+ ],
+ Volatility::Immutable,
+ ),
}
}
}
@@ -154,57 +185,26 @@ impl AggregateUDFImpl for Sum {
&self.signature
}
- fn coerce_types(&self, arg_types: &[DataType]) -> Result<Vec<DataType>> {
- let [args] = take_function_args(self.name(), arg_types)?;
-
- // Refer to
https://www.postgresql.org/docs/8.2/functions-aggregate.html doc
- // smallint, int, bigint, real, double precision, decimal, or interval.
-
- fn coerced_type(data_type: &DataType) -> Result<DataType> {
- match data_type {
- DataType::Dictionary(_, v) => coerced_type(v),
- // in the spark, the result type is
DECIMAL(min(38,precision+10), s)
- // ref:
https://github.com/apache/spark/blob/fcf636d9eb8d645c24be3db2d599aba2d7e2955a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/aggregate/Sum.scala#L66
- DataType::Decimal32(_, _)
- | DataType::Decimal64(_, _)
- | DataType::Decimal128(_, _)
- | DataType::Decimal256(_, _) => Ok(data_type.clone()),
- dt if dt.is_signed_integer() => Ok(DataType::Int64),
- dt if dt.is_unsigned_integer() => Ok(DataType::UInt64),
- dt if dt.is_floating() => Ok(DataType::Float64),
- _ => exec_err!("Sum not supported for {data_type}"),
- }
- }
-
- Ok(vec![coerced_type(args)?])
- }
-
fn return_type(&self, arg_types: &[DataType]) -> Result<DataType> {
match &arg_types[0] {
DataType::Int64 => Ok(DataType::Int64),
DataType::UInt64 => Ok(DataType::UInt64),
DataType::Float64 => Ok(DataType::Float64),
+ // In the spark, the result type is DECIMAL(min(38,precision+10),
s)
+ // ref:
https://github.com/apache/spark/blob/fcf636d9eb8d645c24be3db2d599aba2d7e2955a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/aggregate/Sum.scala#L66
DataType::Decimal32(precision, scale) => {
- // in the spark, the result type is
DECIMAL(min(38,precision+10), s)
- // ref:
https://github.com/apache/spark/blob/fcf636d9eb8d645c24be3db2d599aba2d7e2955a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/aggregate/Sum.scala#L66
let new_precision = DECIMAL32_MAX_PRECISION.min(*precision +
10);
Ok(DataType::Decimal32(new_precision, *scale))
}
DataType::Decimal64(precision, scale) => {
- // in the spark, the result type is
DECIMAL(min(38,precision+10), s)
- // ref:
https://github.com/apache/spark/blob/fcf636d9eb8d645c24be3db2d599aba2d7e2955a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/aggregate/Sum.scala#L66
let new_precision = DECIMAL64_MAX_PRECISION.min(*precision +
10);
Ok(DataType::Decimal64(new_precision, *scale))
}
DataType::Decimal128(precision, scale) => {
- // in the spark, the result type is
DECIMAL(min(38,precision+10), s)
- // ref:
https://github.com/apache/spark/blob/fcf636d9eb8d645c24be3db2d599aba2d7e2955a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/aggregate/Sum.scala#L66
let new_precision = DECIMAL128_MAX_PRECISION.min(*precision +
10);
Ok(DataType::Decimal128(new_precision, *scale))
}
DataType::Decimal256(precision, scale) => {
- // in the spark, the result type is
DECIMAL(min(38,precision+10), s)
- // ref:
https://github.com/apache/spark/blob/fcf636d9eb8d645c24be3db2d599aba2d7e2955a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/aggregate/Sum.scala#L66
let new_precision = DECIMAL256_MAX_PRECISION.min(*precision +
10);
Ok(DataType::Decimal256(new_precision, *scale))
}
diff --git a/datafusion/spark/src/function/aggregate/avg.rs
b/datafusion/spark/src/function/aggregate/avg.rs
index 65736815fe..4a7adc515b 100644
--- a/datafusion/spark/src/function/aggregate/avg.rs
+++ b/datafusion/spark/src/function/aggregate/avg.rs
@@ -15,22 +15,21 @@
// specific language governing permissions and limitations
// under the License.
-use arrow::array::ArrowNativeTypeOp;
use arrow::array::{
builder::PrimitiveBuilder,
cast::AsArray,
types::{Float64Type, Int64Type},
- Array, ArrayRef, ArrowNumericType, Int64Array, PrimitiveArray,
+ Array, ArrayRef, ArrowNativeTypeOp, ArrowNumericType, Int64Array,
PrimitiveArray,
};
use arrow::compute::sum;
use arrow::datatypes::{DataType, Field, FieldRef};
-use datafusion_common::utils::take_function_args;
-use datafusion_common::{not_impl_err, plan_err, Result, ScalarValue};
+use datafusion_common::types::{logical_float64, NativeType};
+use datafusion_common::{not_impl_err, Result, ScalarValue};
use datafusion_expr::function::{AccumulatorArgs, StateFieldsArgs};
use datafusion_expr::utils::format_state_name;
-use datafusion_expr::Volatility::Immutable;
use datafusion_expr::{
- Accumulator, AggregateUDFImpl, EmitTo, GroupsAccumulator, ReversedUDAF,
Signature,
+ Accumulator, AggregateUDFImpl, Coercion, EmitTo, GroupsAccumulator,
ReversedUDAF,
+ Signature, TypeSignatureClass, Volatility,
};
use std::{any::Any, sync::Arc};
@@ -56,7 +55,14 @@ impl SparkAvg {
/// Implement AVG aggregate function
pub fn new() -> Self {
Self {
- signature: Signature::user_defined(Immutable),
+ signature: Signature::coercible(
+ vec![Coercion::new_implicit(
+ TypeSignatureClass::Native(logical_float64()),
+ vec![TypeSignatureClass::Numeric],
+ NativeType::Float64,
+ )],
+ Volatility::Immutable,
+ ),
}
}
}
@@ -66,21 +72,6 @@ impl AggregateUDFImpl for SparkAvg {
self
}
- fn coerce_types(&self, arg_types: &[DataType]) -> Result<Vec<DataType>> {
- let [args] = take_function_args(self.name(), arg_types)?;
-
- fn coerced_type(data_type: &DataType) -> Result<DataType> {
- match &data_type {
- d if d.is_numeric() => Ok(DataType::Float64),
- DataType::Dictionary(_, v) => coerced_type(v.as_ref()),
- _ => {
- plan_err!("Avg does not support inputs of type
{data_type}.")
- }
- }
- }
- Ok(vec![coerced_type(args)?])
- }
-
fn return_type(&self, _arg_types: &[DataType]) -> Result<DataType> {
Ok(DataType::Float64)
}
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
