Data Movement Patterns

How it works

• Create a stream in Load mode with source and target connections
• Schema (tables, keys, indexes, constraints) is converted automatically
• Data is transferred in parallel with per-table progress
• After validation, start CDC for ongoing sync if needed

Join PostgreSQL tables with Parquet files stored in S3 in a single SQL query.

SELECT u.email, o.total
FROM pg.users u
JOIN s3.orders('s3://bucket/orders/*.parquet') o
  ON u.id = o.user_id
WHERE o.total > 100;

Enables ad-hoc analytics across heterogeneous systems instantly.

How it works

• Use Load mode with custom SQL to select a data subset
• Schema is created automatically on the target
• Re-run the same stream whenever you need a fresh copy
• Works across database engines (e.g. production MySQL → dev PostgreSQL)

Reduces drift between production and development systems.