LFC Delta Live Tables 2 — Questions and Answers
Question 1: What is the key difference between a DLT LIVE TABLE and a LIVE VIEW?
- LIVE TABLE stores data persistently; LIVE VIEW is a reusable virtual query that does not store data (Correct answer)
- LIVE TABLE is read-only; LIVE VIEW is writable
- LIVE TABLE uses SQL only; LIVE VIEW uses Python only
- There is no functional difference between them
Correct answer: LIVE TABLE stores data persistently; LIVE VIEW is a reusable virtual query that does not store data
A LIVE TABLE materializes and persists query results as a Delta table, while a LIVE VIEW is a virtual definition reused within the pipeline without storing data.
Question 2: What file format do Delta Live Tables use to store their output by default?
- Parquet
- CSV
- Delta Lake format (Correct answer)
- ORC
Correct answer: Delta Lake format
Delta Live Tables store output in Delta Lake format, enabling ACID transactions, time travel, and schema enforcement on the results.
Question 3: How do you reference another live table within a DLT SQL pipeline?
- Using only the regular table name
- Using the LIVE. prefix before the table name (Correct answer)
- Using a correlated subquery
- Using an external database catalog reference
Correct answer: Using the LIVE. prefix before the table name
Within a DLT pipeline, tables defined in the same pipeline are referenced using the LIVE. namespace prefix (e.g., SELECT * FROM LIVE.my_table).
Question 4: What is a DLT 'pipeline' composed of?
- A single notebook with exactly one table definition
- A collection of notebooks or source files containing dataset definitions and transformations (Correct answer)
- A set of Databricks Jobs linked together
- A cluster configuration YAML file
Correct answer: A collection of notebooks or source files containing dataset definitions and transformations
A DLT pipeline can include one or more notebooks or source files, each contributing table and view definitions that together form the full DAG.
Question 5: Which DLT expectation violation action causes the entire pipeline update to fail?
- WARN
- DROP ROW
- FAIL UPDATE (Correct answer)
- QUARANTINE
Correct answer: FAIL UPDATE
FAIL UPDATE causes the pipeline update to stop with a failure status if any records violate the expectation, enforcing the strictest data quality guarantee.
Question 6: What is Auto Loader in the context of Delta Live Tables pipelines?
- A feature that automatically loads machine learning models into serving
- A tool that incrementally ingests new files arriving in cloud storage into Delta tables (Correct answer)
- A cluster auto-scaling feature for DLT pipelines
- A SQL optimizer that automatically rewrites DLT queries
Correct answer: A tool that incrementally ingests new files arriving in cloud storage into Delta tables
Auto Loader uses file notification or directory listing to efficiently and incrementally ingest new files from cloud storage as they land.
Question 7: What advantage does Continuous pipeline mode offer over Triggered mode?
- Lower compute cost per pipeline run
- Lower latency for near-real-time data processing (Correct answer)
- Simpler pipeline configuration and setup
- Better support for one-time batch workloads
Correct answer: Lower latency for near-real-time data processing
Continuous mode keeps the pipeline running indefinitely and processes new data as it arrives, achieving lower end-to-end latency than Triggered mode.
LFC Delta Live Tables 2
What is the key difference between a DLT LIVE TABLE and a LIVE VIEW?