Design data integration

1. Data from 30 on-premises and SaaS sources must be copied and transformed on a schedule, mostly without code. Which service?

D. Data Factory is the cloud ETL/ELT orchestration service with 90+ connectors.

2. Data Factory must read from a SQL Server inside the corporate network. What do you install?

C. The self-hosted IR runs on-premises (or in a VNet) to reach private data stores.

3. Existing SSIS packages must run in Azure with minimal changes. What do you use?

D. Azure-SSIS IR lifts and shifts SSIS packages.

4. In Data Factory, what defines a unit of work like "copy data" or "run a notebook"?

D. Pipelines group activities; linked services are connections; datasets describe data; triggers start pipelines.

5. What's the main benefit of Azure Data Lake Storage Gen2 for analytics?

C. The hierarchical namespace makes directory operations fast and enables fine-grained ACLs.

6. A data lake is organized as raw, cleansed and curated layers. What's this pattern often called?

B. Data is refined layer by layer from raw (bronze) to business-ready (gold).

7. Data scientists need collaborative notebooks with Apache Spark for machine learning and large-scale transformations. Which service?

C. Databricks is an Apache Spark analytics platform with collaborative notebooks.

8. Which open storage format adds ACID transactions and time travel to data lake files in Databricks?

D. Delta Lake is the default table format in Databricks.

9. A team wants one service for SQL data warehousing, Spark and data pipelines with a single workspace. Which service?

C. Synapse combines dedicated/serverless SQL pools, Spark pools and pipelines.

10. Analysts must query Parquet files in the data lake with T-SQL, paying only per query. Which option?

C. Serverless SQL pools query files in place and bill per TB processed.

11. A large enterprise data warehouse needs predictable performance with MPP. Which Synapse option?

A. Dedicated SQL pools provision compute (DWUs) for massively parallel processing.

12. In a hot/warm/cold data path design, what's the hot path for?

B. Hot path = low latency, e.g. streaming alerts. Cold path = batch over large historical data.

13. Telemetry must trigger alerts within seconds and also be stored for monthly trend analysis. What do you design?

A. This is the lambda-style split: real-time processing plus batch over full history.

14. Which service ingests millions of events per second as the entry point of a streaming pipeline?

A. Event Hubs is the big-data streaming ingestion service (also IoT Hub for device scenarios).

15. Which language does Azure Stream Analytics use to define jobs?

C. Stream Analytics uses a SQL-like language, extensible with JavaScript and C# functions.

16. Average temperature per device must be calculated every 5 minutes, with no overlap between windows. Which window?

B. Tumbling windows are fixed-size, non-overlapping and contiguous.

17. You need a 10-minute average reported every 2 minutes (overlapping windows). Which window?

C. Hopping windows have a size and a hop; they overlap when the hop is smaller than the size.

18. Which are typical Stream Analytics outputs?

C. Stream Analytics can write to many sinks, including real-time Power BI dashboards.

19. Data Factory must start a pipeline whenever a new file lands in a container. Which trigger?

A. Event-based triggers respond to blob created or deleted events via Event Grid.

20. Which service would you choose for low-code data transformations at scale inside Data Factory, without managing Spark?

C. Mapping data flows are visually designed transformations that run on managed Spark clusters.