rothnic
commented
9 years ago I ran into this same problem. Some other options would be:
- Provide a configuration option to proactively alias a specific column in the connection profile
- Check for the conflict, and if it exists prepend the table name
- Put these table properties under their own attribute
db.tables.my_table.meta.count
I implemented the second option in a pull request.
The schema of the database I'm trying to connect to includes a table with column name 'count'. When I try to connect to it, db.py creates a Table object for that table and calls setattr(self, attr, col) for each column in the table. Since Table has a property called 'count', when it gets to that column, it raises an AttributeError. Obviously a similar problem exists with the other method and field names in Table. Other people have mentioned this problem in issue #33
I'm not sure of the best way to handle this. Table.init already checks to make sure we aren't creating a column attribute with name 'name' or 'con'. If so, it puts an underscore in front of the attribute name (this also seems to be a bug, as _con is the name of the field). This behavior isn't necessarily bad, but it does make it hard to find the column names since underscore attributes are kind of hidden in python. For example, in IPython if I have a Table named Foo with a column attribute named _bar, then typing Foo. does not give me _bar as a suggestion. Apart from looking at attribute names on the class, I don't see any way to determine the columns in a table other than db.find_table("Foo").
Perhaps a combination of putting underscores in front of conflicting column name attributes and adding a 'columns' property that would return a list of column names? That still isn't quite ideal, but it would be better than the current situation which completely prevents me from being able to use db.py, which is a shame because it seems pretty cool.
I'd be happy to submit a pull request, but as I said, I'm not sure of the best resolution.