A. Incorrect.
CREATE TABLE is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: SELECT is the table privilege needed to read rows. Broader privileges such as MODIFY or OWN are not required for read-only analytics.

B. Incorrect.
OWN is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: SELECT is the table privilege needed to read rows. Broader privileges such as MODIFY or OWN are not required for read-only analytics.

C. Correct.
SELECT is the table privilege needed to read rows. Broader privileges such as MODIFY or OWN are not required for read-only analytics.

D. Incorrect.
MODIFY is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: SELECT is the table privilege needed to read rows. Broader privileges such as MODIFY or OWN are not required for read-only analytics.

A. Incorrect.
A secret scope is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: Databricks SQL visualizations are created from query results and can be added to dashboards for business consumption.

B. Correct.
Databricks SQL visualizations are created from query results and can be added to dashboards for business consumption.

C. Incorrect.
A cluster policy is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: Databricks SQL visualizations are created from query results and can be added to dashboards for business consumption.

D. Incorrect.
A personal access token is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: Databricks SQL visualizations are created from query results and can be added to dashboards for business consumption.

A. Correct.
Delta time travel lets users query a previous table version or timestamp when history is retained.

B. Incorrect.
OPTIMIZE ZORDER only is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: Delta time travel lets users query a previous table version or timestamp when history is retained.

C. Incorrect.
DROP TABLE IF EXISTS is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: Delta time travel lets users query a previous table version or timestamp when history is retained.

D. Incorrect.
A LEFT ANTI JOIN is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: Delta time travel lets users query a previous table version or timestamp when history is retained.

A. Incorrect.
A dashboard refresh schedule is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: A view stores a reusable SQL query that can simplify repeated transformations and centralize business logic.

B. Incorrect.
A warehouse tag is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: A view stores a reusable SQL query that can simplify repeated transformations and centralize business logic.

C. Incorrect.
A CSV download is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: A view stores a reusable SQL query that can simplify repeated transformations and centralize business logic.

D. Correct.
A view stores a reusable SQL query that can simplify repeated transformations and centralize business logic.

A. Incorrect.
Repos is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: Data Explorer is the Databricks interface for browsing catalogs, schemas, tables, metadata, sample data, lineage, and permissions.

B. Incorrect.
Workflows is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: Data Explorer is the Databricks interface for browsing catalogs, schemas, tables, metadata, sample data, lineage, and permissions.

C. Correct.
Data Explorer is the Databricks interface for browsing catalogs, schemas, tables, metadata, sample data, lineage, and permissions.

D. Incorrect.
Model Serving is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: Data Explorer is the Databricks interface for browsing catalogs, schemas, tables, metadata, sample data, lineage, and permissions.

A. Incorrect.
Removing all filters is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: Auto stop shuts down an idle SQL warehouse after a configured period, reducing compute cost when it is not in use.

B. Correct.
Auto stop shuts down an idle SQL warehouse after a configured period, reducing compute cost when it is not in use.

C. Incorrect.
Increasing result row limits is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: Auto stop shuts down an idle SQL warehouse after a configured period, reducing compute cost when it is not in use.

D. Incorrect.
Disabling query history is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: Auto stop shuts down an idle SQL warehouse after a configured period, reducing compute cost when it is not in use.

A. Correct.
OPTIMIZE with ZORDER co-locates related column values and can improve data skipping for frequent filters.

B. Incorrect.
Changing the dashboard title is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: OPTIMIZE with ZORDER co-locates related column values and can improve data skipping for frequent filters.

C. Incorrect.
Increasing the font size is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: OPTIMIZE with ZORDER co-locates related column values and can improve data skipping for frequent filters.

D. Incorrect.
Dropping all permissions is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: OPTIMIZE with ZORDER co-locates related column values and can improve data skipping for frequent filters.

A. Incorrect.
Owner permissions for everyone is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: View-only permissions let stakeholders consume dashboard results without changing queries, visuals, or dashboard structure.

B. Incorrect.
A public writeable table is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: View-only permissions let stakeholders consume dashboard results without changing queries, visuals, or dashboard structure.

C. Incorrect.
A personal cluster for each viewer is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: View-only permissions let stakeholders consume dashboard results without changing queries, visuals, or dashboard structure.

D. Correct.
View-only permissions let stakeholders consume dashboard results without changing queries, visuals, or dashboard structure.

A. Incorrect.
For a managed table, dropping the table normally removes both table metadata and the managed data files. The files remaining indicates an external table.

B. Incorrect.
A notebook-only library is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: Serverless SQL warehouses provide managed SQL compute with fast startup and reduced infrastructure management for analytics workloads.

C. Correct.
Serverless SQL warehouses provide managed SQL compute with fast startup and reduced infrastructure management for analytics workloads.

D. Incorrect.
A local Spark session is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: Serverless SQL warehouses provide managed SQL compute with fast startup and reduced infrastructure management for analytics workloads.

A. Incorrect.
CROSS JOIN is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: LEFT JOIN keeps all rows from the left table and includes matching rows from the right table when they exist.

B. Correct.
LEFT JOIN keeps all rows from the left table and includes matching rows from the right table when they exist.

C. Incorrect.
INNER JOIN is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: LEFT JOIN keeps all rows from the left table and includes matching rows from the right table when they exist.

D. Incorrect.
LEFT SEMI JOIN only is not the best answer because it does not match the Databricks/Spark behavior being tested. The key point is: LEFT JOIN keeps all rows from the left table and includes matching rows from the right table when they exist.