SqlAlchemy 中操作数据库时session和scoped_session的区别

原生session:

from sqlalchemy.orm import sessionmaker
from sqlalchemy import create_engine
from sqlalchemy应用.models import Users

engine = create_engine(
    "mysql+pymysql://root:[email protected]:3306/pro6?charset=utf8",
    max_overflow=0,  # 超过连接池大小外最多创建的连接
    pool_size=5,  # 连接池大小
)

#from sqlalchemy.orm.session import Session

SessionF = sessionmaker(bind=engine)
session = SessionF()

print(session)

obj1 = Users(name=‘ctz‘, email=‘[email protected]‘, extra=‘aaaa‘)

session.add(obj1)

session.commit()
session.close()

问题：由于无法提供线程共享功能，所以在开发时要注意,要给每个线程都创建自己的session

打印sesion可知他是sqlalchemy.orm.session.Session的对象

查看Session的源码 可得到:

class Session(_SessionClassMethods):
    """Manages persistence operations for ORM-mapped objects.

    The Session‘s usage paradigm is described at :doc:`/orm/session`.

    """

    public_methods = (
        ‘__contains__‘, ‘__iter__‘, ‘add‘, ‘add_all‘, ‘begin‘, ‘begin_nested‘,
        ‘close‘, ‘commit‘, ‘connection‘, ‘delete‘, ‘execute‘, ‘expire‘,
        ‘expire_all‘, ‘expunge‘, ‘expunge_all‘, ‘flush‘, ‘get_bind‘,
        ‘is_modified‘, ‘bulk_save_objects‘, ‘bulk_insert_mappings‘,
        ‘bulk_update_mappings‘,
        ‘merge‘, ‘query‘, ‘refresh‘, ‘rollback‘,
        ‘scalar‘)

2.scoped_session

from sqlalchemy.orm import sessionmaker
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy import create_engine
from sqlalchemy应用.models import Users
from sqlalchemy.orm import scoped_session

engine=create_engine(
    "mysql+pymysql://root:[email protected]:3306/pro6?charset=utf8",
    max_overflow=0,  # 超过连接池大小外最多创建的连接
    pool_size=5,  # 连接池大小
)

SessionF=sessionmaker(bind=engine)

#scoped_session封装了两个值 Session 和 registry,registry加括号就执行了ThreadLocalRegistry的__call__方法,如果

# 当前本地线程中有session就返回session,没有就将session添加到了本地线程

#self.registry()=session
session=scoped_session(SessionF)

print(session)
obj1=Users(name=‘ctz‘,email=‘[email protected]‘,extra=‘aaaa‘)

session.remove()
session.query_property()

优点:支持线程安全,为每个线程都创建一个session:

两种方式:通过本地线程Threading.Local()和创建唯一标识的方法(参考flask请求源码)

源码分析：

首先我们在scoped_session中放入了Sesion对象

SessionF=sessionmaker(bind=engine)

session=scoped_session(Session)

一、scoped_session类中

class scoped_session(object):
    session_factory = None
    def __init__(self, session_factory, scopefunc=None):

        #传递过来的那个Session对象
        self.session_factory = session_factory
        #scopefunc唯一标示函数
        if scopefunc:
            self.registry = ScopedRegistry(session_factory, scopefunc)
        else:
            self.registry = ThreadLocalRegistry(session_factory)

第一次进来scopefunc唯一标识为None，我们将Session作为参数传递到ThreadLocalRegistry中，

ThreadLocalRegistry类中

class ThreadLocalRegistry(ScopedRegistry):
    """A :class:`.ScopedRegistry` that uses a ``threading.local()``
    variable for storage.

    """

    def __init__(self, createfunc):
        #传递过来的那个Session对象
        self.createfunc = createfunc
        self.registry = threading.local()
     #scoped_session.registry()后执行
    def __call__(self):
        try:            #如果本地线程中有值的话直接将值返回，
            return self.registry.value
        except AttributeError:            #没有值的话就示例话Session(),并将他存到本地线程中，并把实例的对象返回
            #相当于Session()后的对象加到了本地线程中
            val = self.registry.value = self.createfunc()
            return val

其中__call__()只有当scoped_session.registry加括号执行

那我们怎么调用那些方法呢？

def instrument(name):
    def do(self, *args, **kwargs):
        return getattr(self.registry(), name)(*args, **kwargs)
    return do

for meth in Session.public_methods:
    setattr(scoped_session, meth, instrument(meth))

其中 Session就是sqlalchemy.orm.session.Session

 public_methods = (
        ‘__contains__‘, ‘__iter__‘, ‘add‘, ‘add_all‘, ‘begin‘, ‘begin_nested‘,
        ‘close‘, ‘commit‘, ‘connection‘, ‘delete‘, ‘execute‘, ‘expire‘,
        ‘expire_all‘, ‘expunge‘, ‘expunge_all‘, ‘flush‘, ‘get_bind‘,
        ‘is_modified‘, ‘bulk_save_objects‘, ‘bulk_insert_mappings‘,
        ‘bulk_update_mappings‘,
        ‘merge‘, ‘query‘, ‘refresh‘, ‘rollback‘,
        ‘scalar‘)

在instrument函数中

self.registry()帮我们执行了ThreadLocalRegistry中的___call__方法，拿到了sesion对象

方法源码示例:

 def has(self):
        return hasattr(self.registry, "value")

    def remove(self):
        if self.registry.has():
            self.registry().close()
        self.registry.clear()

实际两种方式原理都是一样的都是第一种，只是第二种将session放到了本地线程中，为每一个进程都设置了一个session，实现了线程安全

原文地址：https://www.cnblogs.com/ctztake/p/8277372.html