[BP to 6.0.2] - Partition UnnestScan should not use DISTINCT index aggregation (wrong results)

CREATE INDEX ix1 ON default( DISTINCT a1) PARTITION BY HASH(META().id);

SELECT COUNT (DISTINCT ua)

FROM default AS d

UNNEST d.a1 AS ua

WHERE ua IS NOT NULL;

As Index is partition by META().id.
The same value of index keys are spread across different partitions.
As partition keys are not part of group keys Indexer can't determine DISTINCT values
so DISTINCT aggreagtes can't be pushed to indexer.