Revision and Interview Questions

Cheat sheet

Basics: ORM bridges objects and tables; JPA (jakarta.persistence) is the spec, Hibernate the implementation, Spring Data JPA generates repositories; ddl-auto=validate plus Flyway/Liquibase. Entities: @Entity, @Id, no-arg constructor, @Table/@Column, BigDecimal for money, @Enumerated(STRING), embeddables, converters, JSON via @JdbcTypeCode. IDs: SEQUENCE with allocationSize (batch-friendly), IDENTITY (no insert batching), UUID, @NaturalId, @EmbeddedId. Inheritance: @MappedSuperclass, SINGLE_TABLE, JOINED, TABLE_PER_CLASS. Persistence context: transient, managed, detached, removed; persist, find, getReference, merge (returns managed copy), remove; dirty checking; first-level cache; flush before commit and relevant queries. Relationships: @ManyToOne owns the FK; @OneToMany(mappedBy) inverse; helper methods; Set for @ManyToMany; explicit association entity when the link has data; @MapsId one-to-one; cascade and orphanRemoval only for owned children; careful equals/hashCode. Fetching: @ManyToOne/@OneToOne EAGER by default, make LAZY; LazyInitializationException; open-in-view off; N+1 fixed by join fetch, entity graphs, batch fetching, DTO projections. Queries: JPQL with parameters, keyset pagination, Criteria/Specifications for dynamic filters, native SQL, bulk updates bypass the context. Transactions: @Transactional at service layer, rollback on unchecked by default, propagation, self-invocation pitfall; @Version optimistic locking; pessimistic FOR UPDATE; L2 cache for reference data.

Common interview questions

Answer each with an example and a trade-off.

1. What is the difference between JPA, Hibernate and Spring Data JPA?
2. Explain the four entity states and automatic dirty checking.
3. persist vs merge: what does merge return?
4. What is the owning side of a relationship and what does mappedBy do?
5. Why should @ManyToOne be made LAZY, and what is LazyInitializationException?
6. What is the N+1 select problem and how do you fix it?
7. SEQUENCE vs IDENTITY ID generation: which allows batching and why?
8. Compare SINGLE_TABLE and JOINED inheritance strategies.
9. How does @Version prevent lost updates? When would you use a pessimistic lock?
10. Why does @Transactional not work on a method called from the same class?
11. When would you use DTO projections instead of entities?
12. What are the risks of JPQL bulk updates?

Talk in SQL terms

Strong answers explain what SQL Hibernate generates: "join fetch turns 1 + N selects into one join", "@Version adds AND version = ? to the UPDATE". It shows you understand the ORM rather than treating it as magic.

Quick check: Which combination fixes an N+1 problem when listing orders with their customer names?

• Make customer EAGER and enable open-in-view
• Increase the transaction timeout
• Use EnumType.ORDINAL
• Use join fetch or an entity graph for customer, or a DTO projection selecting the name