如何在 SqlAlchemy 中连接不相关的查询

Question

如何按日期连接无关系的 sqlalchemy 查询和 sort？

例如 这些是模型：

Human
Car
Tree

他们都有created 列。以下是查询：

q1 = session.query(Human.created.label('created'), Human).filter(...)
q2 = session.query(Car.created.label('created'), Car).filter(...)
q3 = session.query(Tree.created.label('created'), Tree).filter(...)

现在我想连接这 3 个查询和 order_by 日期。预期的结果会是something，例如：

date      | instance
----------------------------
<created> | Human<instance>
<created> | Car<instance>
<created> | Car<instance>
<created> | Tree<instance>
<created> | Human<instance>
...

Answer 1

根据提供的信息，我猜您想在不同的表上执行联合 (https://www.w3schools.com/sql/sql_union.asp)。

在下面的示例中，创建了不同的表，并将这些不同的表合并在一起。

from sqlalchemy import Column, Integer, String, DateTime
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy import create_engine
from sqlalchemy.orm import sessionmaker
from datetime import datetime

Base = declarative_base()


class Human(Base):
    __tablename__ = 'human'
    id = Column('id', Integer, primary_key=True)
    created = Column('created', DateTime)
    name = Column('name', String(250))

    def __init__(self, created, name):
        self.created = created
        self.name = name

    def __repr__(self):
        return '<{created} - {name}>'.format(created=self.created, name=self.name)


class Car(Base):
    __tablename__ = 'car'
    id = Column('id', Integer, primary_key=True)
    created = Column('created', DateTime)
    brand = Column(String(250))

    def __init__(self, created, brand):
        self.created = created
        self.brand = brand

    def __repr__(self):
        return '<{created} - {brand}>'.format(created=self.created, brand=self.brand)


class Tree(Base):
    __tablename__ = 'tree'
    id = Column('id', Integer, primary_key=True)
    created = Column('created', DateTime)
    type = Column(String(250))

    def __init__(self, created, type):
        self.created = created
        self.type = type

    def __repr__(self):
        return '<{created} - {type}>'.format(created=self.created, type=self.type)


engine = create_engine('sqlite:///')
session = sessionmaker()
session.configure(bind=engine)
ex_ses = session()
Base.metadata.create_all(engine)

human = Human(datetime.now(), 'Human a')
human2 = Human(datetime.now(), 'Human b')
car = Car(datetime.now(), 'Car a')
car2 = Car(datetime.now(), 'Car b')
tree = Tree(datetime.now(), 'Tree a')
tree2 = Tree(datetime.now(), 'Tree b')

ex_ses.add(human)
ex_ses.add(human2)
ex_ses.add(car)
ex_ses.add(car2)
ex_ses.add(tree)
ex_ses.add(tree2)
ex_ses.commit()

# Query
h = ex_ses.query(Human.created, 'name')
c = ex_ses.query(Car.created, 'brand')
t = ex_ses.query(Tree.created, 'type')

print(h.union_all(c).union_all(t).all())

最终查询（在打印语句中）返回：

[(datetime.datetime(2017, 7, 5, 5, 55, 54, 633863), 'Human a'),
(datetime.datetime(2017, 7, 5, 5, 55, 54, 633863), 'Human b'),
(datetime.datetime(2017, 7, 5, 5, 55, 54, 633863), 'Car a'),
(datetime.datetime(2017, 7, 5, 5, 55, 54, 633863), 'Car b'),
(datetime.datetime(2017, 7, 5, 5, 55, 54, 633863), 'Tree a'),
(datetime.datetime(2017, 7, 5, 5, 55, 54, 633863), 'Tree b')]

此答案基于：How to union across multiple tables in SQLAlchemy?