具有以下简化的类和表
tags = db.Table(
"tags",
db.Column("tag_id", db.Integer, db.ForeignKey("tag.id")),
db.Column("post_version_id", db.Integer, db.ForeignKey("post_version.id"))
)
class Tag(db.Model):
id = db.Column(db.Integer, primary_key=True)
class PostVersion(db.Model):
id = db.Column(db.Integer, primary_key=True)
current = db.Column(db.Boolean, default=False, index=True)
我正在尝试通过Tag
的数量排序的所有PostVersion
发出呼叫,这些数量都通过tags
连接并且拥有值current=True
我已经开发了以下查询:
db.session.query(Tag,
db.func.count(PostVersion.id).label("total")
).outerjoin(tags
).outerjoin(PostVersion,
and_(PostVersion.id==tags.c.post_version_id,
PostVersion.current==True)
).group(Tag).order_by("total DESC").all()
产生以下(正确)结果:
[(<Tag 8: original>, 136),
(<Tag 16: constance-garnett>, 136),
(<Tag 3: explanation>, 3),
(<Tag 2: definition>, 1),
(<Tag 14: translation>, 1),
(<Tag 1: biblical>, 0),
(<Tag 4: homage>, 0),
(<Tag 5: intertextuality>, 0),
(<Tag 6: meter>, 0),
(<Tag 7: mythology>, 0),
(<Tag 9: political>, 0),
(<Tag 10: cultural>, 0),
(<Tag 11: reference>, 0),
(<Tag 12: shakespeare>, 0),
(<Tag 13: technical-issues>, 0),
(<Tag 15: context>, 0)]
除了,为了抑制基于元组的输出(其中包含db.func.count(PostVersion.id).label("total")
的结果,我必须对结果进行其他修改,而且我完全不确定如何重写查询来抑制这种情况。
答案 0 :(得分:1)
解决方案很简单:将count表达式移至ORDER BY子句:
db.session.query(Tag).\
outerjoin(tags).\
outerjoin(PostVersion,
and_(PostVersion.id==tags.c.post_version_id,
PostVersion.current==True)).\
group_by(Tag.id).\
order_by(db.func.count(PostVersion.id).desc()).\
all()