Relevant PR comment.

We can likely just enable this for the distinct() command directly, since the distinct command always tries to dedup the output of a DISTINCT_SCAN (even if it doesn't actually have to).

For aggregations, we would also have to disable generating the $groupByDistinct stage (otherwise we may return duplicates for a $group that should only return unique values on its _id/ the DISTINCT_SCAN key), so it is less likely to be useful there.

In either case, doing this would disable SBE usage for $or queries that are distinct-like (until we implement distinct in SBE), so we would need some performance testing to justify whether or not this truly is beneficial.