Possible column name collision

[Former user]

It looks like sqlalchemy gets confused by columns named "something" and "<table_name>_something".

Traceback

  File "/home/jonas/src/sqlalchemy/lib/sqlalchemy/mapping/objectstore.py", line 244, in commit
    get_session().commit(*obj)
  File "/home/jonas/src/sqlalchemy/lib/sqlalchemy/mapping/objectstore.py", line 157, in commit
    self._commit_uow()
  File "/home/jonas/src/sqlalchemy/lib/sqlalchemy/mapping/objectstore.py", line 140, in _commit_uow
    self.uow.commit(*obj)
  File "/home/jonas/src/sqlalchemy/lib/sqlalchemy/mapping/unitofwork.py", line 236, in commit
    commit_context.execute(echo=echo_commit)
  File "/home/jonas/src/sqlalchemy/lib/sqlalchemy/mapping/unitofwork.py", line 353, in execute
    head.execute(self)
  File "/home/jonas/src/sqlalchemy/lib/sqlalchemy/mapping/unitofwork.py", line 507, in execute
    self.mapper.save_obj(self.tosave_objects(), trans)
  File "/home/jonas/src/sqlalchemy/lib/sqlalchemy/mapping/mapper.py", line 666, in save_obj
    c = statement.execute(params)
  File "/home/jonas/src/sqlalchemy/lib/sqlalchemy/sql.py", line 436, in execute
    return c.execute(*multiparams, **params)
  File "/home/jonas/src/sqlalchemy/lib/sqlalchemy/sql.py", line 330, in execute
    return self.engine.execute_compiled(self, params)
  File "/home/jonas/src/sqlalchemy/lib/sqlalchemy/engine.py", line 645, in execute_compiled
    proxy(str(compiled), parameters)
  File "/home/jonas/src/sqlalchemy/lib/sqlalchemy/engine.py", line 640, in proxy    self.execute(statement, parameters, connection=connection, cursor=cursor, return_raw=True)
  File "/home/jonas/src/sqlalchemy/lib/sqlalchemy/engine.py", line 689, in execute
    self._executemany(cursor, statement, parameters)
  File "/home/jonas/src/sqlalchemy/lib/sqlalchemy/engine.py", line 715, in _executemany
    c.executemany(statement, parameters)
pysqlite2.dbapi2.ProgrammingError: Incorrect number of bindings supplied. The current statement uses 2, and there are 1 supplied.
>

Test Case

from sqlalchemy import *

foo = Table('foo',
            Column('id', Integer, primary_key=True),
            Column('foo_id', Integer))

class Foo(object):
    pass

assign_mapper(Foo, foo)
global_connect('sqlite://', echo=True)

foo.create()

f = Foo(foo_id=2)
objectstore.commit()
f.foo_id=3
objectstore.commit()

If I rename the column to "foo_id2" everything works as expected. I'm able to reproduce this with both version 0.1.4 and trunk@1212.

[Mike Bayer]

yeah thats an unfortunate one, since its looking more like column labels should just be more aggressively unique in the first place (i.e. #97)

anyway I checked in the obvious fix which is to add "check for column named the same as the label" to the criterion for creating a randomized label, changeset:1213 , thanks for the catch.