SQLAlchemy使用说明之ORM

对象关系映射(Object Relation Map, ORM)可以将一个类映射为关系模式(数据表). 使用ORM比直接书写SQL在安全性,可读性上都有很大优势.

Working with Related Objects

下面的示例展示如何使用ORM定义一个关系模式并进行实例化.

from sqlalchemy import create_engine, Table, Column, Integer, String, MetaData, ForeignKey
import MySQLdb

from sqlalchemy.orm import sessionmaker
from sqlalchemy.ext.declarative import declarative_base

Base = declarative_base()

class User(Base):
	__tablename__ = 'user'
	user_id = Column('user_id', Integer, primary_key=True)
	name = Column('name', String(20))
	fullname = Column('fullname', String(20))

	def __repr__(self):
		return "<User(user_id='%d', name='%s'), fullname='%s'>" % (self.user_id, self.name, self.fullname)

def get_session():
	engine = create_engine("mysql://root:248536@localhost:3306/test?charset=utf8", encoding="utf-8", echo=True)
	mysql_session_maker = sessionmaker(bind=engine)
	session = mysql_session_maker()
	return session

if __name__ == '__main__':
	session = get_session()
    user = User(user_id=1, name='finley', fullname='finley ?')
    session.add(user)
    session.commit()
    user = User(user_id=2, name='finley2', fullname='finley ?')
    session.add(user)
    session.rollback()
    user = User(user_id=3, name='finley3', fullname='finley ?')
    session.add(user)
    session.commit()
    session.close()

终端回显:

INFO sqlalchemy.engine.base.Engine BEGIN (implicit)
INFO sqlalchemy.engine.base.Engine INSERT INTO user (user_id, name, fullname) VALUES (%s, %s, %s)
INFO sqlalchemy.engine.base.Engine (1, 'finley', 'finley ?')
INFO sqlalchemy.engine.base.Engine COMMIT
INFO sqlalchemy.engine.base.Engine BEGIN (implicit)
INFO sqlalchemy.engine.base.Engine INSERT INTO user (user_id, name, fullname) VALUES (%s, %s, %s)
INFO sqlalchemy.engine.base.Engine (3, 'finley3', 'finley ?')
INFO sqlalchemy.engine.base.Engine COMMIT

上述示例展示了通过ORM和session进行事务操作(implicit transcation).

通过metadata创建表:

Base.metadata.create_all(engine)

rollback()可以回滚到上次commit.

通过session进行事务性操作, 注意那些复杂的工厂.

Add & Remove

生成实例后用session进行添加或者删除操作:

user = User(user_id=1, name='finley', fullname='finley ?')
session.add(user)
session.commit()

删除操作类似, 注意User实例通过查询数据库获得:

user = session.query(User).filter(User.user_id == 1)[0]
session.delete(user)
session.commit()

Query

首先做一个全查询:

result = session.query(User)
session.commit()
session.close()
# print result
print(result)
for instance in result:
	print(instance)

返回的结果:

INFO sqlalchemy.engine.base.Engine   SELECT
user.user_id AS user_user_id,
user.name AS user_name,
user.fullname AS user_fullname
FROM user

INFO sqlalchemy.engine.base.Engine ()
<User(user_id='1', name='finley'), fullname='finley ?'>
<User(user_id='3', name='finley3'), fullname='finley ?'>

也可以很方便的指定要查询的列:

result = session.query(User.user_id, User.name)

结果:

(1L, u'finley')
(3L, u'finley3')

ORM使用filter来筛选记录:

>>> result = session.query(User.user_id, User.name).filter(User.user_id == 1).all()
>>>session.commit()
>>>session.close()
>>>
>>>print(result)
[(1L, u'finley')]

query对象使用和select类似的运算符, filter可以链式调用并支持order_by()等功能.

更多黑魔法请参见

• sqlalchemy docs - ORM - Query

• sqlalchemy docs - ORM - Query with Joins

Relationship

Many to One

Relationship用来维护两个数据表之间的关系, 含有Relationship的表需要使用外键指定参照关系:

from sqlalchemy import create_engine, Table, Column, Integer, String, MetaData, ForeignKey
import MySQLdb

from sqlalchemy.orm import sessionmaker
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy.orm import relationship, backref

Base = declarative_base()

class User(Base):
	__tablename__ = 'user'
	user_id = Column('user_id', Integer, primary_key=True)
	name = Column('name', String(20))
	fullname = Column('fullname', String(20))
	def __repr__(self):
		return "<User(user_id='%d', name='%s', fullname='%s')>" % (self.user_id, self.name, self.fullname)

class Address(Base):
	__tablename__ = 'address'
	address_id = Column('address_id', Integer, primary_key=True)
	user_id = Column(Integer, ForeignKey('user.user_id'))
	user = relationship("User", backref=backref('addresses'))
	def __repr__(self):
		return "<Address(address_id='%d', user_id='%d')>" % (self.address_id, self.user_id)

def get_session():
	engine = create_engine("mysql://root:248536@localhost:3306/test?charset=utf8", encoding="utf-8", echo=True)
	mysql_session_maker = sessionmaker(bind=engine)
	session = mysql_session_maker()
	return session

if __name__ == '__main__':
	session = get_session()
	# add by user
	user = User(user_id=1, name='finley', fullname='finley ?')
	user.addresses = [
		Address(address_id=1, user_id=1),
		Address(address_id=2, user_id=1)
	]
	session.add(user);
	# add by address
	address = Address(address_id=3, user_id=1)
	session.add(address)
	session.commit()
	# query relationship
	result = session.query(User).filter(User.user_id == 1)
	for i in result:
		print(i.addresses)
	session.close()

反向索引backref使得User实例可以查询与其关联的address:

user.addresses

添加关系可以直接操作Address表:

address = Address(address_id=3, user_id=1)
session.add(address) # session.remove(address)
session.commit()

也可以通过User的backref:

    user = User(user_id=1, name='finley', fullname='finley ?')
	user.addresses = [
		Address(address_id=1, user_id=1),
		Address(address_id=2, user_id=1)
	]
	session.add(user);

删除操作类似, 不过因为Foreign Key约束的存在, 需要费点事.

(我不会说自己用删除user记录重新添加的方法的)

Many to Many

sqlalchemy docs - ORM - Many to Many Relationship