零、SQLAlchemy是什么?
SQLAlchemy的官网上写着它的介绍文字:
SQLAlchemy is the Python SQL toolkit and Object Relational Mapper that gives
application developers the full power and flexibility of SQL.
SQLAlchemy 是一个非常强大的ORM和数据库工具,但是它庞大的文档和复杂的功能总是让很 多人望而生畏。而Django的ORM相对来说就让很多人觉得简单实用。
事实上,SQLAlchemy其实也没有那么复杂,光使用它一些比较高级的功能其实并没有比 使用Django ORM复杂多少,而它丰富的功能则能让你在遇到更复杂的问题时处理起来得心应手。
写作本文的主要目的在于:
- 通过对比SQLAlchemy ORM和Django ORM的主要使用方法, 尽量简单直观的让Django用户能够快速了解和上手SQLAlchemy这款强大的工具。
- 不牵扯到SQLAlchemy具体的技术细节,包括Engine连接池、Session的具体工作原理等等
SQLAlchemy相对于Django内建的ORM来说,有几处非常明显的优点:
- 可独立使用,任何使用Python的项目都可以用它来操作数据库
- 和直接使用原始的DBAPI相比,提供了非常丰富的特性:连接池、auto-map等等
- 提供了更底层的SQL抽象语言,能用原始sql解决的问题基本上都可以用SQLAlchemy解决
- 接下来我们针对日常的数据库操作来对比一下Django ORM和SQLAlchemy。
文中使用的 SQLAlchemy 版本为 0.9.8
一、Django VS SQLAlchemy
SQLAlchemy的安装:
1
2
3
4
|
wget http: / / peak.telecommunity.com / dist / ez_setup.py
python ez_setup.py
sudo easy_install sqlalchemy
sudo easy_install ipython
|
1.建立数据表
首先,我们需要先建立几个表。
(1)Django
在Django中,如果要建表,就是在models.py中定义你的数据类型:
1
2
3
4
5
6
7
|
from django.db import models
class Game(models.Model):
... ...
class GameCompany(models.Model):
... ...
|
因为文章主要面向有经验的Django用户,所以此处不写出详细的定义代码。定义Model以后 我们还需要在settings.py中DATABASES处设置需要连接的数据库地址。最后,使用syncdb来 完成数据库表的创建。
(2)SQLAlchemy
在SQLAlchemy中,定义表结构的过程和Django类似:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
|
from sqlalchemy import create_engine
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy import Column, Integer, String, ForeignKey, Date
from sqlalchemy.orm import relationship, backref
Base = declarative_base()
# 定义表结构
class GameCompany(Base):
__tablename__ = 'game_company'
id = Column(Integer, primary_key = True )
name = Column(String( 200 ), nullable = False )
country = Column(String( 50 ))
class Game(Base):
__tablename__ = 'game'
id = Column(Integer, primary_key = True )
company_id = Column(Integer, ForeignKey( 'game_company.id' ), index = True )
category = Column(String( 10 ))
name = Column(String( 200 ), nullable = False )
release_date = Column(Date)
# 和Django不同,外键需要显式定义,具体好坏见仁见智
# 此处的relation可以为lazy加载外键内容时提供一些可配置的选项
company = relationship( 'GameCompany' , backref = backref( 'games' ))
# 此处定义要使用的数据库
engine = create_engine( 'mysql://root:root@localhost:5379/sqlalchemy_tutorial?charset=utf8' )
# 调用create_all来创建表结构,已经存在的表将被忽略
Base.metadata.create_all(engine)
|
2.插入一些数据
接下来,我们往表中插入一些数据
(1)Django
Django中比较常用的插入数据方法就是使用 .save() 了。
1
2
3
4
5
6
7
8
9
10
11
12
|
nintendo = GameCompany(name = "nintendo" , country = "Japan" )
nintendo.save()
game1 = Game(
company = nintendo,
category = "ACT" ,
name = "Super Mario Bros" ,
release_date = '1985-10-18' )
game1.save()
# 或者使用create
Game.objects.create(... ...)
|
(2)SQLAlchemy
在SQLAlchemy ORM中,有一个非常关键的对象 session ,所有对于数据的操作都是 通过session来进行的,所以要插入数据之前,我们得先初始化一个session:
1
2
3
|
from sqlalchemy.orm import sessionmaker
Session = sessionmaker(bind = engine)
session = Session()
|
之后插入数据的方法也和Django比较相似:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
|
# 添加数据
nintendo = GameCompany(name = "Nintendo" , country = "Japan" )
capcom = GameCompany(name = "Capcom" , country = "Japan" )
game1 = Game(
company = nintendo,
category = "ACT" ,
name = "Super Mario Bros" ,
release_date = '1985-10-18'
)
game2 = Game(
company = capcom,
category = "ACT" ,
name = "Devil May Cry 3: Dante's Awakening" ,
release_date = "2005-03-01" ,
)
game3 = Game(
company = nintendo,
category = "RPG" ,
name = "Mario & Luigi: Dream Team" ,
release_date = "2013-08-11" ,
)
# 使用add_all来让这些objects和session产生关系
session.add_all([nintendo, capcom, game1, game2])
# 在没有开启autocommit的模式下,不要忘了调用commit来让数据写到数据库中
session.commit()
|
除了commit之外,session还有rollback()等方法,你可以把session对象简单看成是一次 transaction,所以当你对内容进行修改时,需要调用 session.commit() 来提交这些修改。
去文档可以了解更多session相关内容:http://docs.sqlalchemy.org/en/rel_0_9/orm/session.html
二、常用操作
1.简单查询
(1)批量查询
1
2
3
4
5
6
7
|
# -- Django --
Game.objects. filter (category = "RPG" )
# -- SQLAlchemy --
# 使用filter_by是和django ORM比较接近的方式
session.query(Game).filter_by(category = "RPG" )
session.query(Game). filter (Game.category = = "RPG" )
|
(2)查询单个对象
1
2
3
4
5
6
7
|
# -- Django --
Game.objects.get(name = "Super Mario Bros" )
# -- SQLAlchemy --
session.query(Game).filter_by(name = "Super Mario Bros" ).one()
# `get_objects_or_None()`
session.query(Game).filter_by(name = "Super Mario Bros" ).scalar()
|
Django中得各种 > 、< 都是使用在字段名称后面追加 "__gt"、"__lt" 来实现的,在SQLAlchemy 中这样的查询还要更直观一些
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
|
# -- Django --
Game.objects. filter (release_date__gte = '1999-01-01' )
# 取反
Game.objects.exclude(release_date__gte = '1999-01-01' )
# -- SQLAlchemy --
session.query(Game). filter (Game.release_date > = '1999-01-01' ).count()
# 取反使用 ~ 运算符
session.query(Game). filter (~Game.release_date > = '1999-01-01' ).count()
通过外键组合查询
# -- Django --
Game.objecs. filter (company__name = "Nintendo" )
# -- SQLAlchemy --
session.query(Game).join(GameCompany). filter (GameCompany.name = = "Nintendo" )
|
2.多条件或查询
1
2
3
4
5
6
7
8
|
# -- Django --
from django.db.models import Q
Game.objects. filter (Q(category = "RPG" ) | Q(category = "ACT" ))
# -- SQLAlchemy --
from sqlalchemy import or_
session.query(Game). filter (or_(Game.category = = "RPG" , Game.category = = "ACT" ))
session.query(Game). filter ((Game.category = = "RPG" ) | (Game.category = = "ACT" ))
|
(1)in查询
1
2
3
4
5
|
# -- Django --
Game.objects. filter (category__in = [ "GAL" , "ACT" ])
# -- SQLAlchemy --
session.query(Game). filter (Game.category.in_([ "GAL" , "ACT" ]))
|
(2)like查询
1
2
3
4
5
|
# -- Django --
Game.objects. filter (name__contains = "Mario" )
# -- SQLAlchemy --
session.query(Game.name.contains( 'Mario' ))
|
3.统计个数
简单统计总数:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
|
# -- Django --
Game.objects. filter (category = "RPG" ).count()
# -- SQLAlchemy --
session.query(Game).filter_by(category = "RPG" ).count()
分组统计个数
# -- Django --
from django.db.models import Count
Game.objects.values_list( 'category' ).annotate(Count( 'pk' )).order_by()
# -- SQLAlchemy --
from sqlalchemy import func
session.query(Game.category, func.count(Game.category)).group_by(Game.category). all ()
|
4.结果排序
对查询结果进行排序:
1
2
3
4
5
6
7
8
9
10
11
|
# -- Django --
Game.objects. all ().order_by( 'release_date' )
Game.objects. all ().order_by( '-release_date' )
# 多字段排序
Game.objects. all ().order_by( '-release_date' , 'category' )
# -- SQLAlchemy --
session.query(Game).order_by(Game.release_date)
session.query(Game).order_by(Game.release_date.desc())
# 多字段排序
session.query(Game).order_by(Game.release_date.desc(), Game.category)
|
5.修改数据
1
2
3
4
5
6
7
8
9
|
# -- Django --
game = Game.objects.get(pk = 1 )
game.name = 'Super Mario Brothers'
game.save()
# -- SQLAlchemy --
game = session.query(Game).get( 1 )
game.name = 'Super Mario Brothers'
session.commit()
|
6.批量修改
1
2
3
4
5
|
# -- Django --
Game.objects. filter (category = "RPG" ).update(category = "ARPG" )
# -- SQLAlchemy --
session.query(Game).filter_by(category = "RPG" ).update({ "category" : "ARPG" })
|
7.批量删除
1
2
3
4
5
|
# -- Django --
Game.objects. filter (category = "ARPG" ).delete()
# -- SQLAlchemy --
session.query(Game).filter_by(category = "ARPG" ).delete()
|
三、SQLAlchemy其他一些值得关注的功能
上面简单列了一些SQLAlchemy ORM和Django ORM的使用方法对比,SQLAlchemy同时还提供了一些 其他非常有用的功能,比如Automap~
假如你有一个Django项目,通过ORM创建了一大堆Model。这时来了一个新项目,需要操作 这些表,应该怎么办?拷贝这些Models?使用原始的DB-API加上sql来操作?
其实使用SQLAlchemy的Automap可以让你的工作变得非常的方便,你只要在新项目连接到旧数据库,然后 稍微配置一下Automap,就可以使用SQLAlchemy的ORM操作那些通过别的系统创建的表了。
就像这样:
1
2
3
4
5
6
7
8
9
10
11
|
from sqlalchemy.ext.automap import automap_base
from sqlalchemy.orm import Session
from sqlalchemy import create_engine
Base = automap_base()
engine = create_engine( "sqlite:///mydatabase.db" )
Base.prepare(engine, reflect = True )
# user和address就是表明,通过这样的语句就可以把他们分别映射到User和Address类
User = Base.classes.user
Address = Base.classes.address
|
更多信息可以参考详细文档:http://docs.sqlalchemy.org/en/rel_0_9/orm/extensions/automap.html
附:Django与SQLAlchemy结合的实例演示
譬如,以下gumi/db.py代码,其中gumi制作Django项目名,项目中使用的唯一的数据库连接的包装,作为py调用。
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
|
# -*- coding: utf-8 -*-
from django.conf import settings
from django.core import signals
from django.dispatch import dispatcher
import sqlalchemy
from sqlalchemy.orm import scoped_session, sessionmaker
from sqlalchemy.engine.url import URL
__all__ = [ 'Session' , 'metadata' ]
def create_engine():
url = URL(drivername = settings.DATABASE_ENGINE,
database = settings.DATABASE_NAME,
username = settings.DATABASE_USER,
password = settings.DATABASE_PASSWORD,
host = settings.DATABASE_HOST,
port = settings.DATABASE_PORT or None ,
query = getattr (settings, 'DATABASE_OPTIONS' , {})
)
options = getattr (settings, 'SQLALCHEMY_OPTIONS' , {})
engine = sqlalchemy.create_engine(url, * * options)
return engine
def end_request(signal, sender):
Session.remove()
dispatcher.connect(receiver = end_request,
signal = signals.request_finished)
metadata = sqlalchemy.MetaData()
Session = scoped_session(sessionmaker(autoflush = True ,
transactional = True ,
bind = create_engine()))
|
模块代码
1
2
3
4
5
6
7
8
9
10
|
from sqlalchemy.orm import *
from gumi.db import Session, metadata
some_table = Table( 'some_table' , metadata,
Column( 'id' , Integer, primary_key = True ),
Column( 'some_value' , String( 100 ), nullable = False ,
mysql_engine = 'InnoDB' ,
)
class SomeObject( object ):
pass
mapper(SomeObject, some_table)
|
视图代码
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
|
import django.newforms as forms
from gumi.db import Session
class SomeForm(forms.Form):
# newform
pass
def some_action(req):
if req.method ! = "POST" :
form = SomeForm()
else :
form = SomeForm(req.POST)
if form.is_valid():
data = form.clean()
obj = SomeObject()
obj.some_param = data[ 'a' ]
obj.another_param = data[ 'b' ]
Session.save(obj)
Session.commit()
return HttpResponseRedirect( '/' )
return render_to_response( 'some/template.html' )
|