PostgreSQL connector

Use jdbc:postgresql: for source or sink connections. PostgreSQL supports all three replication modes and bandwidth throttling.

The sink can use the binary COPY path for compatible row shapes. Complex values such as JSON or arrays may use a text-compatible path. Complete-atomic and incremental runs create staging data and then perform the mode-specific merge.

Binary COPY is selected for compatible simple types when a binary value is present. ARRAY, JSON/JSONB, XML, and INTERVAL shapes use TEXT COPY, as do MongoDB source rows whose BSON values do not map directly to PostgreSQL binary encoding. This fallback preserves correctness; it is not an error or a promise of identical throughput for every table shape.

Generated staging tables use the sink table’s shape. Incremental merge depends on a usable sink key, while complete-atomic performs its final replacement in the sink transaction supported by the manager.

Grant the runtime user the table, schema, sequence, and staging permissions required by the chosen mode. Put SSL settings in source.connect.parameter.* or sink.connect.parameter.* and resolve any security values through the environment.