我正在尝试建立跨越四个表的关系。我根据 this question 中的代码简化了我的代码匹配我的数据库。
from sqlalchemy import *
from sqlalchemy.orm import *
from sqlalchemy.ext.declarative import declarative_base
Base = declarative_base()
class A(Base):
__tablename__ = 'a'
id = Column(Integer, primary_key=True)
b_id = Column(Integer, ForeignKey('b.id'))
# FIXME: This fails with:
# "Relationship A.ds could not determine any unambiguous local/remote column pairs based on
# join condition and remote_side arguments. Consider using the remote() annotation to
# accurately mark those elements of the join condition that are on the remote side of the relationship."
#
# ds = relationship("D", primaryjoin="and_(A.b_id == B.id, B.id == C.b_id, D.id == C.d_id)", viewonly=True)
def dq(self):
return sess.query(D).filter(and_(D.id == C.d_id,
C.b_id == B.id,
B.id == A.id,
A.id == self.id))
class B(Base):
__tablename__ = 'b'
id = Column(Integer, primary_key=True)
class C(Base):
__tablename__ = 'c'
b_id = Column(Integer, ForeignKey('b.id'), primary_key=True)
d_id = Column(Integer, ForeignKey('d.id'), primary_key=True)
class D(Base):
__tablename__ = 'd'
id = Column(Integer, primary_key=True)
e = create_engine("sqlite://", echo=True)
Base.metadata.create_all(e)
sess = Session(e)
sess.add(D(id=1))
sess.add(D(id=2))
sess.add(B(id=1))
sess.add(C(b_id=1, d_id=1))
sess.add(C(b_id=1, d_id=2))
sess.add(A(id=1, b_id=1))
sess.flush()
a1 = sess.query(A).first()
print a1.dq().all()
#print a1.ds
所以我的问题是“ds”关系的连接语法。当前错误提到添加 remote(),但我还没有让它工作。我也尝试过使用 secondaryjoin 但没有运气。 “dq”中的查询有效,我最终能够通过在我的代码中使用过滤器来解决它 - 我仍然很好奇如何尽可能构建关系?
最佳答案
我不是 sqlalchemy 专家,这是我的理解。
我认为 sqlalchemy 的关系 API 中的主要混淆来源是,参数 primaryjoin
、secondary
、secondaryjoin
的真正含义是什么。对我来说,它们在这里:
primaryjoin secondaryjoin(optional)
source -------------> secondary -------------------------> dest
(A) (D)
现在我们需要弄清楚中间部分应该是什么。尽管 sqlalchemy 中的自定义连接出乎意料地复杂,但您确实需要了解您的要求,即原始 SQL。一种可能的解决方案是:
SELECT a.*, d.id
FROM a JOIN (b JOIN c ON c.b_id = b.id JOIN d ON d.id = c.d_id) /* secondary */
ON a.b_id = b.id /* primaryjoin */
WHERE a.id = 1;
在这种情况下,源 a
与“辅助”(b JOIN c .. JOIN d ..)
连接,并且没有辅助连接到 D
因为它已经在 secondary
中了。我们有
ds1 = relationship(
'D',
primaryjoin='A.b_id == B.id',
secondary='join(B, C, B.id == C.b_id).join(D, C.d_id == D.id)',
viewonly=True, # almost always a better to add this
)
另一种解决方案可能是:
SELECT a.*, d.id
FROM a JOIN (b JOIN c ON c.b_id = b.id) /* secondary */
ON a.b_id = b.id /* primaryjoin */
JOIN d ON c.d_id = d.id /* secondaryjoin */
WHERE a.id = 1;
这里a
加入secondary(b JOIN c..)
,secondary加入d
与c.d_id = d.id
,因此:
ds2 = relationship(
'D',
primaryjoin='A.b_id == B.id',
secondary='join(B, C, B.id == C.b_id)',
secondaryjoin='C.d_id == D.id',
viewonly=True, # almost always a better to add this
)
经验法则是将长连接路径放在次要路径中,并将其链接到源和目标。
在性能方面,ds1
和 ds2
导致查询计划比 dq
稍微简单一些,但我认为它们之间没有太大区别他们。规划者总是知道得更多。
这是更新后的代码供您引用。请注意如何使用 sess.query(A).options(joinedload('ds1'))
快速加载关系:
from sqlalchemy import *
from sqlalchemy.orm import *
from sqlalchemy.ext.declarative import declarative_base
Base = declarative_base()
class A(Base):
__tablename__ = 'a'
id = Column(Integer, primary_key=True)
b_id = Column(Integer, ForeignKey('b.id'))
ds1 = relationship(
'D',
primaryjoin='A.b_id == B.id',
secondary='join(B, C, B.id == C.b_id).join(D, C.d_id == D.id)',
viewonly=True, # almost always a better to add this
)
ds2 = relationship(
'D',
secondary='join(B, C, B.id == C.b_id)',
primaryjoin='A.b_id == B.id',
secondaryjoin='C.d_id == D.id',
viewonly=True, # almost always a better to add this
)
def dq(self):
return sess.query(D).filter(and_(D.id == C.d_id,
C.b_id == B.id,
B.id == A.id,
A.id == self.id))
class B(Base):
__tablename__ = 'b'
id = Column(Integer, primary_key=True)
class C(Base):
__tablename__ = 'c'
b_id = Column(Integer, ForeignKey('b.id'), primary_key=True)
d_id = Column(Integer, ForeignKey('d.id'), primary_key=True)
class D(Base):
__tablename__ = 'd'
id = Column(Integer, primary_key=True)
def __repr__(self):
return str(self.id)
e = create_engine("sqlite://", echo=True)
Base.metadata.drop_all(e)
Base.metadata.create_all(e)
sess = Session(e)
sess.add(D(id=1))
sess.add(D(id=2))
sess.add(B(id=1))
sess.add(B(id=2))
sess.flush()
sess.add(C(b_id=1, d_id=1))
sess.add(C(b_id=1, d_id=2))
sess.add(A(id=1, b_id=1))
sess.add(A(id=2, b_id=2))
sess.commit()
def get_ids(ds):
return {d.id for d in ds}
a1 = sess.query(A).options(joinedload('ds1')).filter_by(id=1).first()
print('{} a1.ds1: {}'.format('=' * 30, a1.ds1))
assert get_ids(a1.dq()) == get_ids(a1.ds1)
a1 = sess.query(A).options(joinedload('ds2')).filter_by(id=1).first()
print('{} a1.ds2: {}'.format('=' * 30, a1.ds2))
assert get_ids(a1.dq()) == get_ids(a1.ds2)
a2 = sess.query(A).options(joinedload('ds2')).filter_by(id=2).first()
print('{} a2.ds1: {}; a2.ds2 {};'.format('=' * 30, a2.ds1, a2.ds2))
assert a2.ds1 == a2.ds2 == []
关于python - SQLAlchemy 中跨越四个表的关系,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/29302176/