Fastapi sqlalchemy

FastAPI + SQLAlchemy

Patterns for building FastAPI applications with async SQLAlchemy 2.0 and Alembic migrations. Covers session management, repository pattern, relationship loading, transaction handling, and zero-downtime migration strategies.

Core Principles

• Async-first -- Use create_async_engine and async_sessionmaker for non-blocking I/O
• Request-scoped sessions -- One session per request via FastAPI dependency injection
• Repository pattern -- Encapsulate query logic in repository classes, keep routes thin
• Type-safe queries -- Use SQLAlchemy 2.0 select() style instead of legacy Query API
• Small migrations -- Each Alembic revision should make one logical change

When to Use

• Project has sqlalchemy in its dependencies (check pyproject.toml or requirements.txt)
• Building or modifying database models, queries, or migrations
• Setting up async database connectivity in a FastAPI application
• Designing repository or service layers for data access
• Creating or reviewing Alembic migration scripts

Quick Start

Declarative Model

from datetime import datetime

from sqlalchemy import String, func
from sqlalchemy.orm import DeclarativeBase, Mapped, mapped_column


class Base(DeclarativeBase):
    pass


class User(Base):
    __tablename__ = "users"

    id: Mapped[int] = mapped_column(primary_key=True)
    email: Mapped[str] = mapped_column(String(255), unique=True, index=True)
    name: Mapped[str] = mapped_column(String(100))
    is_active: Mapped[bool] = mapped_column(default=True)
    created_at: Mapped[datetime] = mapped_column(server_default=func.now())

Async Engine and Session

from sqlalchemy.ext.asyncio import AsyncSession, async_sessionmaker, create_async_engine

engine = create_async_engine(
    "postgresql+asyncpg://user:pass@localhost/dbname",
    pool_size=20,
    max_overflow=10,
    pool_recycle=3600,
)

async_session_factory = async_sessionmaker(engine, expire_on_commit=False)

FastAPI Session Dependency

from collections.abc import AsyncGenerator

from fastapi import Depends
from sqlalchemy import select


async def get_session() -> AsyncGenerator[AsyncSession, None]:
    async with async_session_factory() as session:
        yield session


@app.get("/users/{user_id}")
async def get_user(
    user_id: int,
    session: AsyncSession = Depends(get_session),
) -> UserResponse | None:
    result = await session.execute(select(User).where(User.id == user_id))
    user = result.scalar_one_or_none()
    return UserResponse.model_validate(user) if user else None

Type Hints

Use | None instead of Optional for all type annotations:

def find_by_email(self, email: str) -> User | None:
    ...

async def list_users(
    self, limit: int = 20, offset: int = 0
) -> list[User]:
    ...

Logging

Use emojis as prefixes in log messages:

logger.info("✅ User %s created successfully", user.id)
logger.warning("⚠️ Slow query detected: %dms for %s", duration_ms, query_name)
logger.error("❌ Database connection failed: %s", exc)

Reference Documents

• sqlalchemy.md -- Async SQLAlchemy 2.0 patterns: declarative models, session lifecycle, query patterns, relationship loading, repository pattern, transactions, and connection pooling
• alembic.md -- Alembic migration management: async setup, autogeneration, data migrations, branching, CI/CD integration, and zero-downtime strategies

Decision Guide

Scenario	Approach
Simple CRUD endpoint	Session dependency + inline query
Complex query logic	Repository class method
Multiple writes in one request	Explicit transaction block
Eager load related data	selectinload for collections, joinedload for scalars
Add a column	Alembic migration with op.add_column
Rename a column safely	Three-step: add new, backfill, drop old
Bulk insert thousands of rows	session.execute(insert(Model), list_of_dicts)
Read-only reporting queries	Separate read replica engine