COUNT(DISTINCT) on Decimal32/Decimal64 doesn’t use the native distinct accumulator #26820

Is your feature request related to a problem or challenge?

COUNT(DISTINCT) on a Decimal32 or Decimal64 column is slower than on a Decimal128 column, even though the values are smaller. In the count_distinct benchmark, a batch of 8,192 Decimal32 values takes about 870 µs, compared with about 110 µs for Decimal128.

get_count_accumulator in datafusion/functions-aggregate/src/count.rs sends Decimal128 and Decimal256 to PrimitiveDistinctCountAccumulator, but has no arm for Decimal32 or Decimal64. They fall through to the generic DistinctCountAccumulator, which converts every row into a ScalarValue.

Describe the solution you'd like

Add Decimal32 and Decimal64 arms to get_count_accumulator that use PrimitiveDistinctCountAccumulator, matching Decimal128 and Decimal256.

Describe alternatives you've considered

No response

Additional context

No response

Willingness to contribute

I plan to contribute to this feature independently