在 SQLAlchemy 中加载连接列的子集答案

【问题标题】：Load subset of joined columns in SQLAlchemy在 SQLAlchemy 中加载连接列的子集
【发布时间】：2016-03-16 04:38:27
【问题描述】：

我正在尝试从 SQLAlchemy 中的多个连接表中加载列的子集，但无法弄清楚神奇的语法。

这是我正在尝试做的一个简单示例。它会导致错误（如下）：

from sqlalchemy import create_engine
from sqlalchemy import Column, Integer, String, ForeignKey
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy.orm import relationship, load_only

Base = declarative_base()

class Table1(Base):
    __tablename__ = 'table1'
    table1_id = Column(Integer, primary_key=True)
    table1_val = Column(String)
    r1 = relationship('Table2', backref = 'r2')


class Table2(Base):
    __tablename__ = 'table2'
    table2_id = Column(Integer, ForeignKey('table1.table1_id'), primary_key=True)
    table2_val = Column(String)


from sqlalchemy.orm import sessionmaker
some_engine = create_engine('postgresql://scott:tiger@localhost/')
Session = sessionmaker(bind=some_engine)
session = Session()
query = session.query(Table1).join(Table2).options(load_only('table1_val','table2_val'))

ArgumentError: Can't find property named 'table2_val' on the mapped entity Mapper|Table1|table1 in this Query.

如何从这些多个表中挑选列？

【问题讨论】：

您是否尝试过使用Table1.table1_val 属性对象，而不仅仅是字符串名称？我希望 SQLAlchemy 在这里需要一个限定名称，因为原则上两个表都可以有重叠的列名。
产生不同的错误：query = session.query(Table1).join(Table2).options(load_only(Table1.table1_val,Table2.table2_val)) ArgumentError: Can't find property 'table2_val' on any entity specified in this Query. Note the full path from root (Mapper|Table1|table1) to target entity must be specified.
另一个变体，另一个错误：query = session.query(Table1,Table2).options(load_only(Table1.table1_val,Table2.table2_val)) 通配符加载器只能用于一个实体。使用 Load(ent) 指定特定实体。
我也尝试过Load() 路由，它不会抛出任何错误，但只加载Table1 中的列，完全忽略Table2 的指定列：query = session.query(Table1).join(Table2).options(Load(Table1).load_only(Table1.table1_val),Load(Table2).load_only(Table2.ta‌ble2_val))
文档 (docs.sqlalchemy.org/en/latest/orm/…) 说这种情况下的语法是 session.query(Table1, Table2).join(Table2).options(Load(Table1).load_only('table1_val'), Load(Table2).load_only('table2_val'))。如果这不起作用，那至少是一个文档错误。

标签： python sqlalchemy

【解决方案1】：

问题已在 [https://bitbucket.org/zzzeek/sqlalchemy/issues/3679/multiple-load_only-in-queryoptions-ignore]

得到解答

以下是更正后的代码：

from sqlalchemy import create_engine
from sqlalchemy import Column, Integer, String, ForeignKey
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy.orm import joinedload, Load, relationship, load_only

Base = declarative_base()

class Table1(Base):
    __tablename__ = 'table1'
    global_id = Column(Integer, primary_key=True)
    table1_val1 = Column(String)
    table1_val2 = Column(String)
    r1 = relationship('Table2', backref = 'r2')

class Table2(Base):
    __tablename__ = 'table2'
    global_id = Column(Integer, ForeignKey('table1.global_id'), primary_key=True)
    table2_val1 = Column(String)
    table2_val2 = Column(String)

from sqlalchemy.orm import sessionmaker
some_engine = create_engine('sqlite://')
Base.metadata.create_all(some_engine)
Session = sessionmaker(bind=some_engine)
session = Session()

session.add(Table1(table1_val1='1val1',table1_val2='1val2',r1=[Table2(table2_val1='2val1', table2_val2='2val2')]))
session.commit() # expires the attribute from the session

query = session.query(Table1).options(
    # note that the 'load_only()' is applied to the 'joinedload' path, not 
    # put on its own Load() path
    joinedload('r1',innerjoin=True).load_only('table2_val1'),
    Load(Table1).load_only('table1_val1'))

foo = query.all()
assert 'table1_val2' not in foo[0].__dict__
assert 'table2_val2' not in foo[0].r1[0].__dict__

【讨论】：