Queries on RLV-enabled tables scan data from both the on-disk IQ main and in-memory RLV stores.
Because of lack of specialized indexes, some query optimizations are not applied for queries on RLV store:
In general, this has a small performance impact, and relates to the proportion of data in the RLV and IQ main stores.
The RLV portion of the query relies on fast in-memory column scans. In some partitioned hash-join cases, in-memory indexes are created on-demand to enhance query performance.