Spring Data & JPA

🗃️ Spring Data & JPA / Hibernate

Core → Advanced

• Entity mapping, relationships (@OneToMany, @ManyToOne, owning side), mappedBy
• Entity lifecycle: transient → managed → detached → removed; persistence context (1st-level cache); dirty checking
• Fetching: LAZY vs EAGER, N+1 problem, JOIN FETCH, @EntityGraph, batch fetching
• Transactions: @Transactional propagation (REQUIRED, REQUIRES_NEW), isolation, read-only, rollback rules
• Locking: optimistic (@Version), pessimistic (PESSIMISTIC_WRITE → SELECT … FOR UPDATE, SKIP LOCKED)
• Projections (interface/DTO/record), pagination (Slice vs Page, keyset)
• Specifications / Querydsl for dynamic queries
• Batch inserts (hibernate.jdbc.batch_size, sequence ID generation)
• Second-level cache (awareness), open-session-in-view (disable it)
• Migrations with Flyway (or Liquibase); expand/contract for zero-downtime deploys
• When to skip JPA: jOOQ or JdbcClient for complex SQL
• HikariCP pool sizing

🧪 Labs (🟢 warm-up → 🟡 core → 🔴 hard → ⚫ boss)

• 🟢 SQL logging on; create an N+1 on workflow → steps; fix it 3 ways
• 🟡 Optimistic locking on workflow publish: 50 concurrent publishes → exactly one wins
• 🔴 Keyset pagination over 5M runs; compare with offset at page 10,000
• 🔴 An outbox table written in the same transaction as the domain change
• ⚫ Rewrite the heaviest query with jOOQ or JdbcClient; compare the SQL and speed

🧠 Cognitive tasks

• Predict the SQL Hibernate will emit for 5 repository methods before running them
• Symptom → hypotheses: “a save() in a loop is slow and memory grows”

🛰️ Orbit integration

• Workflow versions, runs, and prompts repositories in orbit-api

Go deeper

⚙️ Spring Internals · PostgreSQL Internals · 🧩 Enterprise & DDD Patterns

Resources

• Vlad Mihalcea: High-Performance Java Persistence + blog ⭐
• Thorben Janssen’s blog · Spring Data JPA reference