Stream–Table Joins

Joins in motion

• stream–stream join — windowed; you must say within how long.
• stream–table join — point lookup at processing time; the table is your slow-moving dimension (users, products, currencies).
• table–table join — keep two materialised views in sync.

Stream–table is by far the most common: enrich an order event with the current user tier.

Define users_table (a compacted topic), then a continuous query that joins orders_stream to it and emits enriched events.