How do batch processing and stream processing differ, with representative use cases?

How do batch processing and stream processing differ, with representative use cases?

• A. Batch processes large data sets periodically; stream processes continuous data with low latency.
• B. Batch processes continuous data with low latency; stream processes large data sets periodically.
• C. Batch is only for ETL; stream only for analytics.
• D. Batch uses real-time processing; stream uses overnight ETL.

Answer: (A)

The idea being tested is how timing and data velocity shape how processing is done. Batch processing gathers data over a period and then processes it as a unit, which is ideal when you’re handling large volumes and can tolerate a delay between data arrival and results. You might run batch jobs nightly or hourly to perform ETL, update a data warehouse, or retrain a model on historical data. Stream processing, on the other hand, consumes data as it comes in and produces results with very low latency, enabling real-time dashboards, anomaly detection, fraud alerts, or immediate decision-making. So, describing batch as handling large data sets periodically and stream as processing continuous data with low latency matches how these approaches are actually used. The other statements either swap the timing/latency characteristics or overly constrain what batch or stream can do, which doesn’t fit the common practice of batch ETL and real-time analytics.

The idea being tested is how timing and data velocity shape how processing is done. Batch processing gathers data over a period and then processes it as a unit, which is ideal when you’re handling large volumes and can tolerate a delay between data arrival and results. You might run batch jobs nightly or hourly to perform ETL, update a data warehouse, or retrain a model on historical data. Stream processing, on the other hand, consumes data as it comes in and produces results with very low latency, enabling real-time dashboards, anomaly detection, fraud alerts, or immediate decision-making.

So, describing batch as handling large data sets periodically and stream as processing continuous data with low latency matches how these approaches are actually used. The other statements either swap the timing/latency characteristics or overly constrain what batch or stream can do, which doesn’t fit the common practice of batch ETL and real-time analytics.