Intermediate
ETL Staging with Postgres
Build staging layers that tolerate late files and partial retries. Focus on COPY patterns, constraints as data contracts, and quarantine tables that keep support in the loop.
Duration: 4 weeks · 32 hours · Format: Async modules + live clinics
Price (informational): ￥138,000
Request information
What is included
- Staging schema conventions with retention policy
- COPY versus INSERT trade-offs at modest volume
- Quarantine tables with human-readable reasons
- Checksum rows without killing throughput
- Scheduling hooks with observability tags
Outcomes
- Ship a nightly load with explicit validation gates
- Design a quarantine UX support can clear safely
- Document a rollback path for partial file days
Lead contact
Noah Silva
Data platform engineer specializing in regulated imports.
FAQ
We show DAG sketches; orchestrator setup stays in your environment.
CSV and Parquet via foreign data wrappers where applicable.
No Spark tuning; volume assumes single-node Postgres staging.
Participant notes
Quarantine table pattern is now our default for supplier feeds. Mentor pushed on naming — helpful.
Priya