Unifying Tables Across Schemas
A fan-in option merges tables with the same name across schemas or shards into a single destination table, with schema evolution handled automatically.
For teams managing sharded or micro-sharded databases, downstream complexity multiplies fast. Each shard or schema produces its own copy of every table. That means instead of M tables, you end up with N × M tables in your warehouse (where N = number of shards/schemas, M = number of tables). Analysts are stuck stitching them back together, engineers write endless union queries, and operations teams lose the clean, consolidated view they need.
You can now unify tables across schemas directly in replication. Instead of landing one table per schema, Artie automatically merges them into a single, consolidated destination table.
Take an e-commerce platform sharding customers across 50+ schemas: instead of 50 separate users tables, you now get one unified users table downstream. Or a payments company splitting transactions across micro-shards: all those rows flow neatly into one transactions table in Snowflake. With Artie's fan-in option, the number of downstream tables is simplified back to M, and schema evolution is handled automatically.
Why this matters:
- Simplified data model: query one table instead of wrangling dozens, with schema evolution managed for you.
- No duplication: eliminate manual unions or stitching scripts in the warehouse.
- Consistent structure: unified naming across shards improves data quality and usability.
- Effortless scaling: add new shards upstream, and they automatically merge downstream.