181. What is a vector database primarily used for?
- a) Storing and querying high-dimensional vector embeddings for AI applications
- b) Managing geometric spatial data
- c) Storing versioned data
- d) Encrypting data in vector format
Answer: A - Vector databases (e.g., Pinecone, Milvus) specialize in similarity search for ML-generated embeddings.
182. What is the purpose of the SQL ML_MODEL data type?
- a) To store and execute machine learning models within the database
- b) To model database schemas
- c) To create machine learning pipelines
- d) To visualize data distributions
Answer: A - Some databases (e.g., SQL Server) allow storing and scoring ML models directly with PREDICT statements.
183. Which algorithm is commonly used for approximate nearest neighbor (ANN) search in vector databases?
- a) HNSW (Hierarchical Navigable Small World)
- b) B-tree
- c) Dijkstra's algorithm
- d) Bubble sort
Answer: A - HNSW provides efficient approximate similarity search in high-dimensional spaces.
184. What is database DevOps?
- a) Applying CI/CD practices to database schema and code changes
- b) Developing database software
- c) A type of NoSQL database
- d) A backup strategy
Answer: A - Database DevOps uses version control, automated testing, and incremental deployments for databases.
185. What is the purpose of the SQL VECTOR_EMBEDDING data type?
- a) To store dense numerical arrays for AI similarity operations
- b) To implement geometric vectors
- c) To encrypt data in vector form
- d) To model graph edges
Answer: A - Emerging SQL extensions support vector embeddings (e.g., pgvector in PostgreSQL).
186. What is automated database tuning?
- a) Using AI/ML to optimize configurations without human intervention
- b) Automatically restructuring schemas
- c) A type of database backup
- d) Self-repairing databases
Answer: A - Systems like Oracle Autonomous Database or Azure SQL Database auto-tune indexes and queries.
187. What is the purpose of the SQL SEMANTIC_SIMILARITY function?
- a) To measure similarity between text embeddings or vectors
- b) To compare database schemas
- c) To validate semantic versions
- d) To implement fuzzy joins
Answer: A - Semantic similarity functions (e.g., cosine similarity) enable "find similar items" queries.
188. What is a cloud-native database?
- a) A database designed for horizontal scaling and microservices architectures
- b) Any database hosted in the cloud
- c) A database with web interfaces
- d) A serverless computing model
Answer: A - Cloud-native databases (e.g., CockroachDB) leverage elasticity, APIs, and distributed systems principles.
189. What is the purpose of the SQL GENERATE_AI_EMBEDDING function?
- a) To transform text/images into vector embeddings using built-in models
- b) To create AI-generated data
- c) To train machine learning models
- d) To visualize vector spaces
Answer: A - Some databases now integrate embedding generation (e.g., OpenAI embeddings in Supabase).
190. What is database observability?
- a) Comprehensive monitoring of performance, health, and usage patterns
- b) Making databases readable
- c) A backup verification technique
- d) A security audit method
Answer: A - Observability combines metrics, logs, and traces to understand database behavior.
191. What is the purpose of the SQL VECTOR_SEARCH operator?
- a) To find rows with similar vector embeddings
- b) To search through vector graphics
- c) To optimize geometric queries
- d) To implement full-text search
Answer: A - Vector search enables "find similar" queries for recommendation systems and AI applications.
192. What is a feature store in machine learning databases?
- a) A centralized repository for curated ML features
- b) A database of product features
- c) A type of vector database
- d) A cloud storage system
Answer: A - Feature stores (e.g., Feast) manage consistent feature datasets for training and inference.
193. What is the purpose of the SQL AI_TRAIN function?
- a) To train machine learning models using database data
- b) To optimize query training
- c) To educate database administrators
- d) To generate training datasets
Answer: A - Some databases (e.g., BigQuery ML) enable model training with SQL syntax.
194. What is database continuous integration (CI)?
- a) Automatically testing schema changes against application code
- b) Merging databases continuously
- c) A type of replication
- d) A backup synchronization method
Answer: A - Database CI validates migrations, constraints, and performance before deployment.
195. What is the purpose of the SQL VECTOR_INDEX type?
- a) To accelerate similarity searches on vector columns
- b) To index vector graphics
- c) To implement geometric indexes
- d) To compress vector data
Answer: A - Vector indexes (e.g., IVF, HNSW) optimize nearest-neighbor queries in embedding spaces.
196. What is a database GitOps workflow?
- a) Managing database changes through Git pull requests and automated pipelines
- b) Using Git as a database
- c) A version control system for backups
- d) A collaborative editing system
Answer: A - GitOps applies software development practices to database schema management.
197. What is the purpose of the SQL AI_SCORE function?
- a) To apply ML model inference to query results
- b) To rate database performance
- c) To evaluate AI accuracy
- d) To benchmark queries
Answer: A - AI scoring functions enable real-time predictions within SQL queries.
198. What is database infrastructure as code (IaC)?
- a) Defining database configurations in machine-readable definition files
- b) Writing stored procedures in general-purpose languages
- c) A type of NoSQL database
- d) Automatically generating database code
Answer: A - IaC tools like Terraform manage database instances, users, and permissions declaratively.
199. What is the purpose of the SQL VECTOR_AGGREGATE function?
- a) To perform mathematical operations on vector columns
- b) To combine vector graphics
- c) To implement geometric aggregations
- d) To compress multiple vectors
Answer: A - Vector aggregates enable operations like "average embedding" for clustering.
200. What is a serverless database?
- a) A database that automatically scales compute based on demand
- b> A database without a server component
- c> A single-user database
- d> An in-memory database
Answer: A - Serverless databases (e.g., Aurora Serverless, Firebase) abstract capacity management with pay-per-use pricing.