Databricks Certified Data Engineer - Associate

Explanation

Correct Answer: A (Database)

Why Database is the correct choice:

1. Persistent Storage: A database in Databricks is a logical collection of tables that is stored in a physical location (typically in Unity Catalog or the Hive metastore). It persists across sessions and can be accessed by multiple users.

2. Cross-session Accessibility: Databases are catalog objects that can be accessed by other data engineers in different sessions, as they are registered in the metastore.

3. Physical Storage: Databases have associated physical storage locations where table data is actually stored (in cloud storage like S3, ADLS, etc.).

4. Table Aggregation: A database can contain multiple tables, which aligns with the requirement of creating a data entity "from a couple of tables."

Why Function is incorrect:

1. Not a Data Entity: Functions (UDFs - User Defined Functions) are code objects, not data entities. They are used to transform data, not to store or organize data.

2. No Physical Storage: Functions don't have physical storage locations for data; they are stored as code definitions in the metastore.

3. Different Purpose: Functions are used for data processing and transformation, not for organizing and persisting data structures.

Additional Context: In Databricks, databases (also called schemas in some contexts) are the primary way to organize tables logically. They provide:

• Namespace isolation
• Access control at the database level
• Physical storage management
• Metadata management through Unity Catalog

The data engineer should create a database to organize the tables, which will then be accessible to other engineers across different sessions and will have physical storage backing.