sqalchemy update bindparam主键

Question

以下代码抛出“sqlalchemy.exc.CompileError错误：未使用的列名：\ id“。你知道吗

User = Table('users', metadata,
 Column('id', Integer, primary_key=True),
 Column('score', Integer)
)

values = [
    {'score': 2, '_id': 1},
    {'score': 3, '_id': 3}
]
query = User.update().where(User.c.id == bindparam('_id')).values(score=bindparam('score'))
await db.execute_many(query, values)

db是databases.Database的实例。注意，我必须使用名称'\u id'，因为SQLalchemy说'id'是保留的。你知道吗

除了个别地更新每一行之外，还有别的解决办法吗？你知道吗

Answer 1

Database.execute_many()调用Connection.execute_many()，它将您的查询分解为单独的查询（在values中每个元素一个），下面是方法（source）：

    async def execute_many(
        self, query: typing.Union[ClauseElement, str], values: list
    ) -> None:
        queries = [self._build_query(query, values_set) for values_set in values]
        async with self._query_lock:
            await self._connection.execute_many(queries)

注意，它调用_build_query()方法（source）：

    @staticmethod
    def _build_query(
        query: typing.Union[ClauseElement, str], values: dict = None
    ) -> ClauseElement:
        if isinstance(query, str):
            query = text(query)

            return query.bindparams(**values) if values is not None else query
        elif values:
            return query.values(**values)

        return query

当您不传递str查询并且传递值时，control进入elif values:条件处理，其中值的单个dict被解压到查询的.values()方法中（即^{}）。这使得它试图编译的查询：

query = (
    User.update()
    .where(User.c.id == bindparam("_id"))
    .values(score=bindparam("score"))
    .values(score=2, _id=1)
)

第二个values子句产生一个新的更新，新的bind参数试图设置score和_id的值。这会导致查询编译失败，因为表上没有_id列。你知道吗

所以要重现错误的MCVE实际上是这样的：

from sqlalchemy.dialects import postgresql

User.update().values(score=2, _id=1).compile(dialect=postgresql.dialect())

这引发了：

Traceback (most recent call last):
  File ".\main.py", line 31, in <module>
    User.update().values(score=2, _id=1).compile(dialect=postgresql.dialect())
  File "<string>", line 1, in <lambda>
  File "C:\Users\peter\Documents\git\stackoverflow\58668615-sqalchemy-update-bindparam-primary-key\.venv\lib\site-packages\sqlalchemy\sql\elements.py", line 462, in compile
    return self._compiler(dialect, bind=bind, **kw)
  File "C:\Users\peter\Documents\git\stackoverflow\58668615-sqalchemy-update-bindparam-primary-key\.venv\lib\site-packages\sqlalchemy\sql\elements.py", line 468, in _compiler
    return dialect.statement_compiler(dialect, self, **kw)
  File "C:\Users\peter\Documents\git\stackoverflow\58668615-sqalchemy-update-bindparam-primary-key\.venv\lib\site-packages\sqlalchemy\sql\compiler.py", line 571, in __init__
    Compiled.__init__(self, dialect, statement, **kwargs)
  File "C:\Users\peter\Documents\git\stackoverflow\58668615-sqalchemy-update-bindparam-primary-key\.venv\lib\site-packages\sqlalchemy\sql\compiler.py", line 319, in __init__
    self.string = self.process(self.statement, **compile_kwargs)
  File "C:\Users\peter\Documents\git\stackoverflow\58668615-sqalchemy-update-bindparam-primary-key\.venv\lib\site-packages\sqlalchemy\sql\compiler.py", line 350, in process
    return obj._compiler_dispatch(self, **kwargs)
  File "C:\Users\peter\Documents\git\stackoverflow\58668615-sqalchemy-update-bindparam-primary-key\.venv\lib\site-packages\sqlalchemy\sql\visitors.py", line 92, in _compiler_dispatch
    return meth(self, **kw)
  File "C:\Users\peter\Documents\git\stackoverflow\58668615-sqalchemy-update-bindparam-primary-key\.venv\lib\site-packages\sqlalchemy\sql\compiler.py", line 2569, in visit_update
    self, update_stmt, crud.ISUPDATE, **kw
  File "C:\Users\peter\Documents\git\stackoverflow\58668615-sqalchemy-update-bindparam-primary-key\.venv\lib\site-packages\sqlalchemy\sql\crud.py", line 62, in _setup_crud_params
    return _get_crud_params(compiler, stmt, **kw)
  File "C:\Users\peter\Documents\git\stackoverflow\58668615-sqalchemy-update-bindparam-primary-key\.venv\lib\site-packages\sqlalchemy\sql\crud.py", line 177, in _get_crud_params
    % (", ".join("%s" % c for c in check))
sqlalchemy.exc.CompileError: Unconsumed column names: _id

为了总结这个问题，您构建了一个查询，其中bind参数同时传递给Update.where()和Update.values()。然后将该查询和值传递给Database.execute_many()，在那里它们将值列表中的各个元素解压到查询上的第二个Update.values()调用中，该调用将您的查询替换为试图为不存在的_id列设置值的查询。你知道吗

Is there any solution other than updating each row individullay?

当使用sqlalchemy引擎和查询时，查询工作正常：

# using a sqlalchemy engine
engine.execute(query, values)

否则，应该将查询作为字符串发送到Database.execute_many()，因为这意味着查询将在_build_query()方法的if isinstance(query, str):部分处理，这将避免对查询进行第二次.values()调用：

db.execute_many(str(query), values)