Defines a custom aggregation level for deduplication.
This is used to dedupe over a different aggregation level than the default behavior. For example, aggregation over an intermediate table or another key. This can be used for similar use cases to level-of-detail calculation.