refactor: DecimalSumState overflow handling in SUM aggregation (#17773)

fix: DecimalSumState overflow handling in SUM aggregation

Fix the boolean logic in the implementation to correctly handle decimal overflow
checking in aggregate SUM operations.

Rename the generic parameter from OVERFLOW to SHOULD_CHECK_OVERFLOW to better
reflect its purpose.

Author: dantengsky

src/query/functions/src/aggregates/aggregate_sum.rs

@@ -165,7 +165,7 @@ where
 }
 
 #[derive(BorshDeserialize, BorshSerialize)]
-pub struct DecimalSumState<const OVERFLOW: bool, T>
+pub struct DecimalSumState<const SHOULD_CHECK_OVERFLOW: bool, T>
 where
     T: ValueType,
     T::Scalar: Decimal,
@@ -174,7 +174,7 @@ where
     pub value: <T::Scalar as Decimal>::U64Array,
 }
 
-impl<const OVERFLOW: bool, T> Default for DecimalSumState<OVERFLOW, T>
+impl<const SHOULD_CHECK_OVERFLOW: bool, T> Default for DecimalSumState<SHOULD_CHECK_OVERFLOW, T>
 where
     T: ValueType,
     T::Scalar: Decimal + std::ops::AddAssign,
@@ -187,7 +187,8 @@ where
     }
 }
 
-impl<const OVERFLOW: bool, T> UnaryState<T, T> for DecimalSumState<OVERFLOW, T>
+impl<const SHOULD_CHECK_OVERFLOW: bool, T> UnaryState<T, T>
+    for DecimalSumState<SHOULD_CHECK_OVERFLOW, T>
 where
     T: ValueType,
     T::Scalar: Decimal + std::ops::AddAssign,
@@ -201,7 +202,7 @@ where
         let mut value = T::Scalar::from_u64_array(self.value);
         value += T::to_owned_scalar(other);
 
-        if OVERFLOW && (value > T::Scalar::MAX || value < T::Scalar::MIN) {
+        if SHOULD_CHECK_OVERFLOW && (value > T::Scalar::MAX || value < T::Scalar::MIN) {
             return Err(ErrorCode::Overflow(format!(
                 "Decimal overflow: {:?} not in [{}, {}]",
                 self.value,
@@ -219,7 +220,7 @@ where
         validity: Option<&Bitmap>,
         function_data: Option<&dyn FunctionData>,
     ) -> Result<()> {
-        if !OVERFLOW {
+        if !SHOULD_CHECK_OVERFLOW {
             let mut sum = T::Scalar::from_u64_array(self.value);
             let col = T::upcast_column(other);
             let buffer = DecimalType::<T::Scalar>::try_downcast_column(&col).unwrap();
@@ -362,20 +363,20 @@ pub fn try_create_aggregate_sum_function(
             };
 
             // DecimalWidth<int64_t> = 18
-            let overflow = s.precision > 18;
+            let should_check_overflow = s.precision > 18;
             let return_type = DataType::Decimal(DecimalDataType::from_size(decimal_size)?);
 
-            if overflow {
+            if should_check_overflow {
                 AggregateUnaryFunction::<
-                    DecimalSumState<false, Decimal128Type>,
+                    DecimalSumState<true, Decimal128Type>,
                     Decimal128Type,
                     Decimal128Type,
                 >::try_create_unary(
                     display_name, return_type, params, arguments[0].clone()
                 )
             } else {
                 AggregateUnaryFunction::<
-                    DecimalSumState<true, Decimal128Type>,
+                    DecimalSumState<false, Decimal128Type>,
                     Decimal128Type,
                     Decimal128Type,
                 >::try_create_unary(
@@ -399,20 +400,20 @@ pub fn try_create_aggregate_sum_function(
                 scale: s.scale,
             };
 
-            let overflow = s.precision > 18;
+            let should_check_overflow = s.precision > 18;
             let return_type = DataType::Decimal(DecimalDataType::from_size(decimal_size)?);
 
-            if overflow {
+            if should_check_overflow {
                 AggregateUnaryFunction::<
-                    DecimalSumState<false, Decimal256Type>,
+                    DecimalSumState<true, Decimal256Type>,
                     Decimal256Type,
                     Decimal256Type,
                 >::try_create_unary(
                     display_name, return_type, params, arguments[0].clone()
                 )
             } else {
                 AggregateUnaryFunction::<
-                    DecimalSumState<true, Decimal256Type>,
+                    DecimalSumState<false, Decimal256Type>,
                     Decimal256Type,
                     Decimal256Type,
                 >::try_create_unary(