Data Engineering Systems: 10 practice questions

Question 1 of 10

What is the central responsibility of data engineering?

1. AOperate transaction screens while avoiding analytical data products
2. BStore every byte without making it accessible to any consumer
3. CCreate business forecasts without preparing or governing source data
4. DBuild systems that deliver reliable, usable data for downstream work

Question 2 of 10

Which activity belongs to an operational workload?

1. AScanning integrated sales data to prepare an executive dashboard
2. BComparing historical revenue trends across product categories
3. CGrouping five years of orders by region and quarter
4. DRecording and updating a customer's current order

Question 3 of 10

What is the role of ingestion in a data platform?

1. AMove batch or streaming data from sources into the platform
2. BCoordinate task dependencies without moving source data
3. CDecide the business meaning of every analytical measure before any source is ingested
4. DExpose final dashboards without landing or processing data

Question 4 of 10

Which responsibility belongs to compute rather than storage?

1. AKeep raw objects available between separate pipeline runs
2. BMaintain the persistent files that form a table version
3. CRetain data durably after the processing job stops
4. DExecute a join and aggregation over stored records

Question 5 of 10

Which responsibility belongs to orchestration?

1. APerform the join merely by storing its schedule definition without running transformation compute
2. BStart dependent tasks in the required order and manage retries
3. CPersist every source file as the compute engine's memory
4. DDefine product category as a numeric sales measure

Question 6 of 10

How does metadata make a data product more usable?

1. AIt replaces the data values with task schedules and retry counters
2. BIt makes storage unnecessary after the asset is cataloged
3. CIt supplies schema, lineage, ownership, and access context for the asset
4. DIt performs every transformation without requiring compute resources

Question 7 of 10

Why separate analytical processing from a transactional database?

1. AAnalytical systems must update each order before the operational system records it
2. BLarge analytical scans can run without competing directly with current transaction handling
3. CTransactional systems cannot store current records or enforce reliable updates
4. DSeparation prevents analytical systems from using historical data from several sources

Question 8 of 10

A dashboard is stale because its transformation never started after ingestion finished. The raw data is present and the transformation code works when run manually. Which component should be corrected first?

1. AOrchestration, because the dependency and task start were not coordinated
2. BMetadata, because lineage should perform the missing aggregation itself
3. CStorage, because durable raw data should execute transformation code automatically
4. DServing, because the dashboard should reconstruct raw data without transformation

Question 9 of 10

A checkout service needs fast current-order updates, while finance needs monthly trends across orders, payments, and refunds. Which separation best fits both workloads?

1. AKeep each source isolated and ask finance to infer trends from current records only
2. BRun all monthly scans on the checkout database during peak transaction traffic
3. CKeep transaction updates operational and serve integrated history from an analytical repository
4. DMove checkout writes into a dashboard dataset and remove the operational record store

Question 10 of 10

Users can query a table, but they cannot tell its owner, freshness, source, or permitted audience. Compute and storage are healthy. Which investment addresses the actual gap?

1. AAdd more transformation compute even though queries already execute successfully
2. BAdd a scheduler that reruns the table continuously without defining ownership
3. CImprove catalog, lineage, ownership, freshness, and access metadata
4. DDuplicate the same table in more storage locations without descriptions