Module Assessment

1.

Your AI application needs to store conversation metadata with varying structures. Some conversations include custom fields like model configuration, while others have user preferences. Which PostgreSQL data type best supports this requirement?

TEXT

JSONB

VARCHAR(MAX)

2.

You insert a new conversation record and need the auto-generated ID immediately to insert related messages. Which PostgreSQL clause retrieves the generated value without a separate query?

RETURNING

OUTPUT

SELECT LAST_INSERT_ID()

3.

You need to insert a user preference record. If the record already exists, you want to update the value without creating a duplicate. Which PostgreSQL clause handles this scenario?

INSERT INTO table ON DUPLICATE KEY UPDATE

INSERT INTO table ON CONFLICT DO UPDATE

INSERT INTO table IF NOT EXISTS

4.

Your Python application creates many short-lived database connections to store individual messages during AI inference. What approach should you use to improve performance?

Open a new connection for each message and close it immediately after

Use a ConnectionPool to maintain reusable connections that the application can borrow and return

Keep a single global connection open for the entire application lifetime

5.

You're designing a table for AI agent task checkpoints. Tasks have a status that should only be 'pending', 'in_progress', 'completed', or 'failed'. Which constraint enforces this rule at the database level?

UNIQUE (status)

CHECK (status IN ('pending', 'in_progress', 'completed', 'failed'))

NOT NULL DEFAULT 'pending'

Feedback