Reviewed by Aditya Kumar · Last reviewed 2026-03-24
Architectural components: PCollection (bounded/unbounded dataset), PTransform (Map, Filter, GroupByKey), Pipeline (graph), Runner (Dataflow). Side inputs for lookups; watermarks for event-time. Flow: ReadFromPubSub → Map → WindowInto → GroupByKey → WriteToBigQuery. Why Beam:...
Architectural components: PCollection (bounded/unbounded dataset), PTransform (Map, Filter, GroupByKey), Pipeline (graph), Runner (Dataflow). Side inputs for lookups; watermarks for event-time. Flow: ReadFromPubSub → Map → WindowInto → GroupByKey → WriteToBigQuery. Why Beam: Portable; batch and stream. Best practice: Structured data; design for parallelism; allowed lateness for late data.
Red Flag: Ignoring late data handling in streaming. Pro-Move: 'We use allowedLateness=1h and side input for enrichment—backlog cleared, no late data loss.'
According to DataEngPrep.tech, this is one of the most frequently asked Cloud/Tools interview questions, reported at 1 company. DataEngPrep.tech maintains an editor-reviewed database of 1,863 data engineering interview questions across 7 categories.