Add JOIN support for pure parallel replicas

Why it matters

This feature enables efficient and consistent JOIN queries on tables replicated purely in parallel, improving query parallelism and reliability in distributed setups where replicas are independently processing parts of the data.

How to use it

Users can leverage this feature by executing JOIN clauses on tables replicated with pure parallel replicas. No additional configuration is required beyond setting up the distributed tables and replicas according to existing parallel replication methods.