django-orm

The Django Object-Relational Mapper (ORM) is a layer that abstracts database interactions, allowing developers to query and manipulate relational database tables using Python code and object-oriented paradigms instead of writing raw SQL.

SQL syntax can differ between database engines (PostgreSQL, SQLite, MySQL) and writing raw SQL strings in Python is error-prone. The ORM acts as a translator. You write Python methods like .filter(), and the ORM translates that into the correct, optimized SQL for your specific database.

1. Queries are constructed using a model’s Manager (e.g., objects).
2. Method calls (filter, exclude, annotate) return a QuerySet.
3. QuerySets are lazy: they don’t hit the database until they are evaluated (e.g., iterated over, sliced, or cast to a list).
4. Upon evaluation, the ORM generates SQL and executes it against the database backend.

graph TD
  A[Python: Model.objects.filter()] --> B(Django ORM Translator)
  B --> C[SQL: SELECT * FROM table WHERE...]
  C --> D[(Database)]

Think of the ORM as a personal translator in a foreign country. You speak your native language (Python) to the translator, and the translator seamlessly speaks the local language (SQL) to the locals (the Database), handling all the grammatical nuances.

• Lazy evaluation for efficiency.
• Database agnostic (switch DBs without changing query code).
• Protects against SQL injection automatically.

• Built from: Django Model — models are the core of the ORM.
• Builds into: Django Migration — schemas are managed by the ORM state.
• Related: Django View — views use the ORM to fetch data.
• Related: Python — leverages Python’s OOP.

• “N+1 query problem”: Accessing related objects in a loop without using select_related or prefetch_related causes severe performance issues.
• Very complex analytical queries might still require raw SQL for performance reasons.