What are two common data ingestion patterns in Fabric, and when would you use them?

• A. Batch ingestion via Dataflow or pipelines for periodic loads; streaming/near-real-time via spark-based streaming or event-driven pipelines for continuous data updates
• B. Batch ingestion via REST and SQL pipelines; streaming via event hubs only
• C. Ingestion via manual copy only
• D. Ingestion via Graph queries only

Answer: (A)

Two common data ingestion patterns in Fabric are batch ingestion and streaming ingestion. Batch ingestion loads data in discrete, scheduled batches using Dataflow or pipelines, which is ideal for large, periodic data loads from source systems where a small delay is acceptable—for example nightly or hourly ETL from transactional systems. Streaming ingestion brings data continuously as it arrives, enabling near-real-time updates, and can be implemented with Spark-based streaming (Structured Streaming) or event-driven pipelines for immediate analysis, such as real-time dashboards, telemetry, or logs.

The other options don’t match how Fabric typically handles data intake. REST and SQL pipelines aren’t the standard paired patterns for ongoing data ingestion, manual copy isn’t scalable for regular loads, and graph queries aren’t an ingestion pattern.