What are the features in Snowflake? How do you implement CDC in Snowflake?

Snowflake is a cloud‑native data warehouse that offers several key features:

1. Separation of storage and compute – you can scale each independently.
2. Zero‑copy cloning – create instant copies of databases, schemas, or tables without duplicating data.
3. Time Travel – query historical data up to 90 days and recover dropped objects.
4. Continuous Data Protection (CDP) – automatically back up data to a separate region.
5. Native support for semi‑structured data – JSON, Avro, Parquet, XML can be queried with SQL.
6. Automatic scaling and concurrency – virtual warehouses spin up on demand.

Implementing CDC in Snowflake typically uses Streams and Tasks:

• Create a stream on the source table: CREATE OR REPLACE STREAM my_stream ON TABLE my_table APPEND_ONLY = FALSE; The stream captures INSERT, UPDATE, and DELETE changes.
• Create a task that runs a SQL statement to read from the stream and write changes to a target table or a change log.
• The task can be scheduled (e.g., every minute) or set to run continuously.
• Optionally, use Snowpipe for real‑time ingestion of change logs from external sources.

This approach provides near real‑time CDC with minimal overhead and leverages Snowflake’s built‑in change tracking.