Struct datafusion_physical_expr_common::aggregate::utils::DecimalAverager

pub struct DecimalAverager<T: DecimalType> { /* private fields */ }

Computes averages for Decimal128/Decimal256 values, checking for overflow

This is needed because different precisions for Decimal128/Decimal256 can store different ranges of values and thus sum/count may not fit in the target type.

For example, the precision is 3, the max of value is 999 and the min value is -999

Implementations

impl<T: DecimalType> DecimalAverager<T>

pub fn try_new( sum_scale: i8, target_precision: u8, target_scale: i8, ) -> Result<Self>

Create a new DecimalAverager:

• sum_scale: the scale of sum values passed to Self::avg
• target_precision: the output precision
• target_scale: the output scale

Errors if the resulting data can not be stored

pub fn avg(&self, sum: T::Native, count: T::Native) -> Result<T::Native>

Returns the sum/count as a i128/i256 Decimal128/Decimal256 with target_scale and target_precision and reporting overflow.

• sum: The total sum value stored as Decimal128 with sum_scale (passed to Self::try_new)
• count: total count, stored as a i128/i256 (NOT a Decimal128/Decimal256 value)