Batch and Streaming Ingestion Patterns: 10 practice questions

Question 1 of 10

Which execution mode represents batch ingestion in the documented ingestion schedule?

1. ATriggered mode, which remains active and processes each record at arrival
2. BContinuous mode, which can begin only through a manual pipeline start
3. CContinuous mode, which waits for a schedule before checking for new data
4. DTriggered mode, which processes new data on a schedule or after a manual start

Question 2 of 10

What defines continuous ingestion mode?

1. AIt processes the full source again whenever a schedule starts
2. BIt requires a manual trigger before every group of source records
3. CIt processes new data only at the end of a recurring schedule
4. DIt processes new data as the data arrives in the source

Question 3 of 10

Which capability belongs to Apache Spark Structured Streaming?

1. AEnd-to-end fault tolerance and exactly-once processing through Spark APIs
2. BSource-specific authentication and long-term API maintenance without Spark APIs
3. CScheduled-only processing with no support for arrival-driven execution
4. DA declarative layer limited to scheduled batch ingestion

Question 4 of 10

What does Lakeflow pipelines add when extending Structured Streaming?

1. ADeclarative management available only for batch pipelines
2. BSource-specific authentication and long-term API maintenance for every source
3. CA requirement that every pipeline run continuously instead of on a schedule
4. DDeclarative management of orchestration, monitoring, data quality, and errors

Question 5 of 10

For a supported source, what operational work does the fully-managed connector layer take over?

1. AAutomated retries, schema evolution, and source-specific authentication
2. BManual retries, fixed schemas, and user-maintained source authentication
3. CSpark API control without source-specific handling or maintenance
4. DScheduled execution that disables change data capture and edge-case handling

Question 6 of 10

How should the ingestion implementation layer normally be selected?

1. AStart with continuous mode and move to triggered mode when more customization is needed
2. BStart with managed connectors and keep them even when the source is unsupported
3. CStart with the most managed layer and move down if it cannot meet the requirement
4. DStart with Structured Streaming and move up when more automation is required

Question 7 of 10

Why does selecting Lakeflow pipelines not determine the ingestion cadence?

1. ALakeflow pipelines select a source but cannot define transformations or execution
2. BLakeflow pipelines run only after manual triggers but accept both SQL and Python
3. CLakeflow pipelines run continuously but switch languages for scheduled workloads
4. DLakeflow pipelines can build both batch and streaming pipelines in SQL and Python

Question 8 of 10

A pipeline must process the files accumulated during each business day at 02:00. Which execution mode fits this requirement?

1. AContinuous mode because daily files require arrival-driven processing
2. BTriggered mode because processing should begin on a recurring schedule
3. CTriggered mode because it must remain active between scheduled runs
4. DContinuous mode because it should wait until the scheduled start time

Question 9 of 10

A team needs arrival-driven processing and direct control through Spark APIs. Which combination best fits?

1. ATriggered execution with Lakeflow pipelines because arrivals require schedules
2. BContinuous execution with a managed connector for maximum API customization
3. CTriggered execution with a fully-managed connector for direct Spark API control
4. DContinuous execution with Apache Spark Structured Streaming

Question 10 of 10

A supported streaming source needs arrival-driven processing, declarative transformations, monitoring, and data-quality management. Which design is most suitable?

1. ATriggered execution through managed connectors because source arrivals form a batch
2. BContinuous execution through Structured Streaming because declarative management is unavailable
3. CTriggered execution through Structured Streaming because monitoring requires schedules
4. DContinuous execution through Lakeflow pipelines