ORM vs raw SQL

ORMs map classes to tables and methods to queries—productive for CRUD. Raw SQL wins for complex reports, bulk operations, and performance-critical paths where you need full control.

ORM benefits

• Less boilerplate for simple CRUD
• Type-safe models and migrations integration
• Database abstraction across dev SQLite and prod Postgres

Raw SQL benefits

SELECT date_trunc('month', ordered_at) AS month,
       SUM(total) AS revenue
FROM orders
GROUP BY 1
ORDER BY 1;

Dialect-specific functions, window queries, and EXPLAIN tuning are often clearer in raw SQL—common in Data Science notebooks via pandas/SQLAlchemy text().

Hybrid approach

Use ORM for app transactions; drop to raw SQL for admin dashboards, migrations, and batch ETL. Log and review generated SQL in development.

Important interview questions and answers

1. Q: ORM lazy loading?
   A: Related objects fetched on access—can cause N+1 unless eager loaded.
2. Q: When skip ORM?
   A: Heavy aggregates, bulk COPY, or queries needing specific index hints/plans.

Self-check

1. One advantage of ORM for CRUD?
2. One case where raw SQL is preferable?

Tip: Log ORM-generated SQL in dev; switch to raw SQL for complex reports when needed.