I have clickhouse cluster with distributed tables sharded by Id. I need to select data from distributed table 'A_distributed' that is joined locally with 'B_local' tables on Id and then given result without globals joins. It is possible in theory because same ids are will be on the same shard because of distributed key. Is that possible in clickhouse?
Look https://clickhouse.com/docs/sql-reference/statements/select/join#distributed-join
actually this is how exactly clickhouse work, if you don't use GLOBAL JOIN or GLOBAL IN .. then JOIN will proceed on destination nodes with underlying local table for A_distributed and local table B_local