is data sharded consistently across databases with the same number of partitions on the same number of nodes such that a query would result in local join operations, or is sharding consistency only guaranteed within the same database? For example, suppose that database A has table X with shard key K and database B has table Y with the same shard key K, then I join them with the following query.
SELECT *
FROM A.X
INNER JOIN B.Y
ON Y.K = X.K;