Python 数据库实战：SQLite、PostgreSQL、SQLAlchemy 与 ORM 完全指南

Thu, 13 Aug 2026 16:00:00 +0800

几乎所有 Python 应用都需要持久化数据。本文从 SQLite 起步，到 SQLAlchemy ORM，再到异步数据库访问，给你一个完整的数据库技术栈。

1. 数据库选型指南

数据库	适用场景	Python 驱动
SQLite	单机应用、测试、嵌入式	sqlite3（内置）
PostgreSQL	生产应用、复杂查询	psycopg2, asyncpg
MySQL	Web 应用、LAMP 栈	PyMySQL, mysqlclient
MongoDB	文档型数据	pymongo
Redis	缓存、消息队列	redis-py

2. SQLite：零配置数据库

2.1 基础 CRUD

import sqlite3
from pathlib import Path

DB_PATH = Path("app.db")

def init_db():
 with sqlite3.connect(DB_PATH) as conn:
 conn.execute("""
 CREATE TABLE IF NOT EXISTS users (
 id INTEGER PRIMARY KEY AUTOINCREMENT,
 name TEXT NOT NULL,
 email TEXT UNIQUE,
 created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
 )
 """)

def add_user(name: str, email: str):
 with sqlite3.connect(DB_PATH) as conn:
 conn.execute("INSERT INTO users (name, email) VALUES (?, ?)", (name, email))
 conn.commit()

def get_users():
 with sqlite3.connect(DB_PATH) as conn:
 conn.row_factory = sqlite3.Row # 返回字典式结果
 return conn.execute("SELECT * FROM users").fetchall()

# 使用
init_db()
add_user("Alice", "alice@example.com")
for user in get_users():
 print(dict(user))

2.2 SQLite 高级特性

# 事务控制
with sqlite3.connect(DB_PATH) as conn:
 try:
 conn.execute("INSERT INTO users (name) VALUES (?)", ("Bob",))
 conn.execute("INSERT INTO invalid_table ...") # 会失败
 conn.commit()
 except sqlite3.Error:
 conn.rollback()

# 返回字典结果
conn.row_factory = lambda c, r: {col[0]: r[idx] for idx, col in enumerate(c.description)}

# FTS5 全文搜索（需要编译时启用）
conn.execute("CREATE VIRTUAL TABLE docs USING fts5(title, content)")

3. PostgreSQL 与 psycopg2

# pip install psycopg2-binary
import psycopg2
from psycopg2.extras import RealDictCursor

conn = psycopg2.connect(
 host="localhost",
 database="mydb",
 user="myuser",
 password="mypass",
)

with conn.cursor(cursor_factory=RealDictCursor) as cur:
 cur.execute("SELECT * FROM users WHERE age > %s", (18,))
 for row in cur.fetchall():
 print(dict(row))

conn.close()

4. SQLAlchemy ORM 核心

4.1 模型定义

# pip install sqlalchemy
from sqlalchemy import create_engine, Column, Integer, String, DateTime, ForeignKey
from sqlalchemy.orm import declarative_base, relationship, sessionmaker
from datetime import datetime

Base = declarative_base()

class User(Base):
 __tablename__ = "users"
 
 id = Column(Integer, primary_key=True)
 name = Column(String(100), nullable=False)
 email = Column(String(255), unique=True)
 created_at = Column(DateTime, default=datetime.utcnow)
 
 posts = relationship("Post", back_populates="author")

class Post(Base):
 __tablename__ = "posts"
 
 id = Column(Integer, primary_key=True)
 title = Column(String(200), nullable=False)
 content = Column(String)
 author_id = Column(Integer, ForeignKey("users.id"))
 
 author = relationship("User", back_populates="posts")

# 创建表
engine = create_engine("sqlite:///app.db", echo=True)
Base.metadata.create_all(engine)

# Session
Session = sessionmaker(bind=engine)

4.2 CRUD 操作

# 创建
with Session() as session:
 user = User(name="Alice", email="alice@example.com")
 session.add(user)
 session.commit()
 print(user.id) # 自动获取自增 ID

# 查询
with Session() as session:
 # 主键查询
 user = session.get(User, 1)
 
 # 条件查询
 users = session.query(User).filter(User.name.like("A%")).all()
 
 # 分页
 page = session.query(User).offset(10).limit(10).all()

# 更新
with Session() as session:
 user = session.get(User, 1)
 user.name = "New Name"
 session.commit()

# 删除
with Session() as session:
 user = session.get(User, 1)
 session.delete(user)
 session.commit()

5. 关系与查询

5.1 关联查询

# 一对多：用户 - 文章
with Session() as session:
 user = session.get(User, 1)
 for post in user.posts:
 print(post.title)

# 预加载（避免 N+1 查询）
from sqlalchemy.orm import joinedload

users = session.query(User).options(joinedload(User.posts)).all()

# 聚合查询
from sqlalchemy import func

result = session.query(
 User.name,
 func.count(Post.id).label("post_count")
).join(Post).group_by(User.id).all()

6. 异步数据库：asyncpg

# pip install asyncpg
import asyncpg
import asyncio

async def main():
 conn = await asyncpg.connect("postgresql://user:pass@localhost/db")
 
 # 查询
 rows = await conn.fetch("SELECT * FROM users WHERE age > $1", 18)
 for row in rows:
 print(row["name"])
 
 # 事务
 async with conn.transaction():
 await conn.execute("INSERT INTO users (name) VALUES ($1)", "Bob")
 
 await conn.close()

asyncio.run(main())

SQLAlchemy 异步版本

from sqlalchemy.ext.asyncio import create_async_engine, AsyncSession
from sqlalchemy.orm import sessionmaker

async_engine = create_async_engine("postgresql+asyncpg://user:pass@localhost/db")
AsyncSessionLocal = sessionmaker(async_engine, class_=AsyncSession)

async def get_users():
 async with AsyncSessionLocal() as session:
 result = await session.execute(select(User).where(User.age > 18))
 return result.scalars().all()

7. 数据库迁移：Alembic

# 安装
pip install alembic

# 初始化
alembic init migrations

# 修改 alembic.ini → sqlalchemy.url = postgresql://user:pass@localhost/db
# 修改 migrations/env.py → 导入你的 Base

# 生成迁移脚本
alembic revision --autogenerate -m "create users table"

# 升级
alembic upgrade head

# 降级
alembic downgrade -1

8. 生产环境最佳实践

使用连接池（SQLAlchemy 内置）
敏感信息存环境变量
索引外键列
定期备份（pg_dump、mysqldump）
查询慢日志分析

SQLAlchemy on PlumePHP

Python 数据库实战：SQLite、PostgreSQL、SQLAlchemy 与 ORM 完全指南

目录

1. 数据库选型指南

2. SQLite：零配置数据库

2.1 基础 CRUD

2.2 SQLite 高级特性

3. PostgreSQL 与 psycopg2

4. SQLAlchemy ORM 核心

4.1 模型定义

4.2 CRUD 操作

5. 关系与查询

5.1 关联查询

6. 异步数据库：asyncpg

SQLAlchemy 异步版本

7. 数据库迁移：Alembic

8. 生产环境最佳实践

延伸阅读