[sqlalchemy] Re: Modify the start number of a sequence after it's been created

2009-12-16 Thread Jason R. Coombs
Here's what I do. First, I get the sequence f rom the table's columns seq = table.c.id.default # or if you have a mapper-class # seq = MyClass.id_property.property.columns[0].default Then, I execute that sequence to advance the sequence until it's the value needed. engine.execute(seq) The

[sqlalchemy] Problem with long identifiers for indexes generated by SQLAlchemy in Oracle 9i

2007-11-29 Thread Jason R. Coombs
Hello. I've run into a problem with sqlalchemy 0.3.12dev (svn trunk) on Oracle 9i where I get an exception identifier too long. This is on table creation where the the column is Column( 'signer_accountid', index=True ) and the tablename is 'signed_documents'. The index generated is

[sqlalchemy] Re: Multiple inserts and last_inserted_ids

2007-11-27 Thread Jason R. Coombs
This issue seems to be causing me some trouble as well. Using sqlalchemy-0.4.2dev_r3832, I'm getting the following error when attempting to flush my session: if not len(self._last_inserted_ids) or self._last_inserted_ids[0] is None: AttributeError: 'MSSQLExecutionContext_pyodbc' object has