Databricks Interview prep - Unity Catalog & Governance

Question 1

What is Unity Catalog?

Answer 1

Unity Catalog is a centralized governance layer for data and AI assets:
* Manages access control across workspaces
* Provides data lineage
* Enables fine-grained security (row/column level)
👉 It replaces the legacy Hive Metastore with enterprise-grade governance.

Question 2

What is the object hierarchy in Unity Catalog?

Answer 2

Catalog → Schema → Table/View
Example:
catalog = finance
schema = transactions
table = payments
👉 This structure supports multi-tenant environments.

Question 3

What are the key differences between Unity Catalog and Hive Metastore?

Answer 3

Unity Catalog:
Centralized across workspaces
Fine-grained access control
Built-in lineage

Hive Metastore:
Workspace-level
Limited governance
No native lineage
👉 Unity Catalog is designed for enterprise governance at scale.

Question 4

What types of access control does Unity Catalog support?

Answer 4

Table-level
Column-level
Row-level filtering
👉 Enables:
Data masking
Compliance (PII protection)

Question 5

How does row-level security work in Unity Catalog?

Answer 5

• Applies filters based on user identity
• Users see only permitted rows
  👉 Example:
  Sales manager sees only their region

Question 6

How is column-level security implemented?

Answer 6

• Mask or restrict access to specific columns
  👉 Example:
• Hide salary or SSN for non-authorized users

Question 7

What is data lineage and why is it important?

Answer 7

Tracks:
* Data origin
* Transformations
* Downstream dependencies

👉 Helps with:
* Debugging
* Impact analysis
* Compliance audits

Question 8

What is an external location in Unity Catalog?

Answer 8

• A secured reference to cloud storage (e.g., ADLS, S3)
  👉 Controls access to:
• Files
• Tables stored outside managed storage

Question 9

What are storage credentials in Unity Catalog?

Answer 9

• Secure authentication mechanism to access cloud storage
  👉 Separates:
• Access control from compute
• Credentials from code

Question 10

What is the difference between managed and external tables?

Answer 10

• Managed: Databricks controls data + metadata
• External: Data stored outside, only metadata managed 👉 Managed = easier 👉 External = more control

Question 11

How does Unity Catalog enable data sharing?

Answer 11

• Secure sharing across workspaces or organizations
• No need to duplicate data 👉 Often uses Delta Sharing.

Question 12

What is the principle of least privilege and why is it important?

Answer 12

• Users get only the access they need 👉 Reduces:
• Security risks
• Accidental data exposure

Question 13

How is RBAC implemented in Unity Catalog?

Answer 13

• Assign permissions to roles/groups
• Grant roles to users 👉 Easier to manage than user-level permissions.

Question 14

What is data masking and when should it be used?

Answer 14

• Obscures sensitive data
  👉 Example:
• Show only last 4 digits of credit card Used for: * Compliance (GDPR, HIPAA)

Question 15

How does Unity Catalog handle multiple workspaces?

Answer 15

• Central governance across all workspaces
  👉 Ensures:
• Consistent policies
• No duplication of rules

Question 16

Why is auditing important in data governance?

Answer 16

• Tracks who accessed what data and when 👉 Required for:
• Compliance
• Security monitoring

Question 17

Why is defining data ownership important?

Answer 17

• Ensures accountability
• Improves data quality
• Clarifies responsibility
  👉 Each dataset should have a clear owner.

Question 18

What is the tradeoff between governance and flexibility?

Answer 18

• Strong governance → secure but restrictive
• High flexibility → agile but risky
  👉 Balance is key in enterprise environments.

Question 19

How would you secure PII data in Databricks?

Answer 19

• Use column-level masking
• Apply row-level filters
• Restrict access via RBAC
  👉 Combine multiple controls for strong security.

Question 20

Why is it important to understand data access patterns?

Answer 20

• Helps define permissions
• Improves performance
• Avoids overexposure
  👉 Governance should align with usage.

Question 21

How do you enforce governance in data pipelines?

Answer 21

• Apply permissions at table level
• Validate data quality
• Track lineage
  👉 Governance is not just access control—it’s end-to-end.

Question 22

What are common mistakes in data governance?

Answer 22

• Over-permissioning users
• No data ownership
• Ignoring lineage
• Poor documentation
  👉 Leads to “data chaos”.

Question 23

How does Unity Catalog integrate with Delta Lake?

Answer 23

• Delta provides data reliability
• Unity Catalog provides governance
  👉 Together:
• Reliable + secure data platform

Question 24

How would you handle a request for sensitive data access?

Answer 24

• Validate business need
• Apply least privilege
• Grant temporary or scoped access
  👉 Always audit and monitor usage.

Question 25

What does a good governance strategy look like in Databricks?

Answer 25

* Centralized control (Unity Catalog) * Fine-grained permissions * Data lineage tracking * Clear ownership 👉 Goal: * Secure + scalable + compliant data platform