4

I am running SQLAlchemy against FirebirdSQL, and when I execute an insert command in my project, SQLAlchemy is raising an exception on returning from executing against the connection. However, the insert query is being constructed and executed successfully. Querying the database shows that the items are actually being inserted correctly.

Edit: I'm digging down into the fbcore.py module now, and checking the value of value and vartype indicates that the issue is probably how the SEQUENCE item used to generate the primary key ID is returning its data is at issue. The vartype is SQL_LONG, but the actual value is [<an integer>] where <an integer> is the value returned by a sequence generator I created to auto-increment the primary key (e.g. [14]). This suggests to me that the problem should be resolved by fixing that, though I'm not sure how to do it. The generator appears to be working correctly within the database itself, but causing problems when returned to SQLAlchemy.

See below for my existing implementation and the stack trace for details.

My code:

class Project:
    # (I've snipped project instantiation, where engine connection, table, etc. are configured)
    def save_project(self, id_=None, title=None, file_name=None, file_location=None):

        # Build the dictionary of values to store
        values = {}
        if title is not None:
            values['title'] = title

        if file_name is not None:
            values['file_name'] = file_name

        if file_location is not None:
            values['file_location'] = file_location

        # Simplification: I account for the case that there *is* data---skipping that here

        # Execute the correct kind of statement: insert or settings_update.
        if id_ is None:
            statement = self.table.insert()

        else:
            statement = self.table.update().where(self.table.c.id == id_)

        result = self.connection.execute(statement, values)

        # If we inserted a row, get the new primary key. Otherwise, return
        # the one specified by the user; it does not change on settings_update.
        project_id = result.inserted_primary_key if result.is_insert else id_

The traceback:

  File "/Users/chris/development/quest/workspace/my_project/data/tables.py", line 350, in save_project
    result = self.connection.execute(statement, values)
  File "/Users/chris/.virtualenvs/my_project/lib/python3.3/site-packages/sqlalchemy/engine/base.py", line 720, in execute
    return meth(self, multiparams, params)
  File "/Users/chris/.virtualenvs/my_project/lib/python3.3/site-packages/sqlalchemy/sql/elements.py", line 317, in _execute_on_connection
    return connection._execute_clauseelement(self, multiparams, params)
  File "/Users/chris/.virtualenvs/my_project/lib/python3.3/site-packages/sqlalchemy/engine/base.py", line 817, in _execute_clauseelement
    compiled_sql, distilled_params
  File "/Users/chris/.virtualenvs/my_project/lib/python3.3/site-packages/sqlalchemy/engine/base.py", line 947, in _execute_context
    context)
  File "/Users/chris/.virtualenvs/my_project/lib/python3.3/site-packages/sqlalchemy/engine/base.py", line 1111, in _handle_dbapi_exception
    util.reraise(*exc_info)
  File "/Users/chris/.virtualenvs/my_project/lib/python3.3/site-packages/sqlalchemy/util/compat.py", line 168, in reraise
    raise value
  File "/Users/chris/.virtualenvs/my_project/lib/python3.3/site-packages/sqlalchemy/engine/base.py", line 940, in _execute_context
    context)
  File "/Users/chris/.virtualenvs/my_project/lib/python3.3/site-packages/sqlalchemy/dialects/firebird/kinterbasdb.py", line 106, in do_execute
    cursor.execute(statement, parameters or [])
  File "/Users/chris/.virtualenvs/my_project/lib/python3.3/site-packages/fdb/fbcore.py", line 3323, in execute
    self._ps._execute(parameters)
  File "/Users/chris/.virtualenvs/my_project/lib/python3.3/site-packages/fdb/fbcore.py", line 2991, in _execute
    self.__Tuple2XSQLDA(self._in_sqlda, parameters)
  File "/Users/chris/.virtualenvs/my_project/lib/python3.3/site-packages/fdb/fbcore.py", line 2782, in __Tuple2XSQLDA
    sqlvar.sqlscale)
  File "/Users/chris/.virtualenvs/my_project/lib/python3.3/site-packages/fdb/fbcore.py", line 2266, in _check_integer_range
    if (value < vmin) or (value > vmax):
TypeError: unorderable types: list() < int()

I'm not yet sufficiently familiar with SQLAlchemy's to see why this is an issue; the style of my statement is pretty much identical to that in the tutorial. This appears to be an issue with how the parameters are being passed – possibly something about using a dict rather than keyword arguments? But there's nothing in the docs on how to handle parameters that suggests I have anything amiss here – it seems right from what I'm seeing there.

I've also tried this with self.table.insert().values(values) rather than passing the values term to the execute method, with the same results (as I would expect).

Edit: I note from reading the docstring on execute in fbcore.py that it raises a TypeError when the parameters passed to the method are not given either as a list or a tuple. Is this a change that is not yet reflected in the documentation?

Edit 2: As a comment notes, the stack trace indicates that it's running against the kinterbasdb driver, though I have explicitly configured the engine to run using fdb. This is also confusing to me.

4

2 回答 2

3

正如我所预料的那样，特别是当我发现问题是该行按预期插入但UPDATE不久之后用函数调用时，问题出在一些相关代码上。我将结果返回为project_id（如您在上面的代码中所见），并且由于完全不相关的原因（与 Blinker 信号有关），该方法再次被调用，返回值为project_id，我已设置如下：

project_id = result.inserted_primary_key if result.is_insert else id_

此行的正确版本仅略有不同：

project_id = result.inserted_primary_key[0] if result.is_insert else id_

来自SQLAlchemy 文档（强调我的）：

返回刚刚插入的行的主键。

返回值是与目标表中的主键列列表相对应的标量值列表。

这里的返回值必须是一个列表，因为主键可以是数据库中多个字段的组合。（这对我来说应该很明显；很明显我已经一年多没有做过认真的数据库工作了。）由于这种情况下的主键是单个值，所以我只是选择了那个值并返回它，问题是解决。

当然，现在我必须去寻找 Blinker 信号问题——这个方法不应该被调用两次——但是 c'est la vie...

于 2014-04-04T20:39:20.087 回答
1

我一直在查看 SQL Alchemy 文档，我想知道您是否应该这样做：

if id_ is None:
    statement = self.table.insert()

else:
    statement = self.table.update().where(self.table.c.id == id_)

statement = statement.values(title=title, file_name=file_name, file_location=file_location)

result = self.connection.execute(statement)

也就是说：不要将字典传递给执行，而是将其作为语句的一部分（如Insert Expressions所示）。

于 2014-04-04T16:04:39.370 回答