Cannot use NamedTupleRow about pypyodbc

[GoogleCodeExporter]

I have been stepping through the source code trying to fix this issue myself, 
and so far I have worked out the following:

1. The Row object is dynamically created with the list of fields (using the 
recordtype module)
2. This field list is empty, it gets populated from the _Col
3. When I run a query using MutableNamedTupleRow, Cursor._ColBufferList does 
not get correctly populated. This means that the Row object does not get 
created properly.

Is this likely to be fixed any time soon? I don't really know enough about the 
low-level portions of the pypyodbc code to fix the _ColBufferList.

It might be possible to use the Cursor.description list to create the Row 
object instead; I'm not sure.

Original comment by [email protected] on 21 May 2014 at 7:52

[GoogleCodeExporter]

I believe that what I said in my previous comment is correct; 
Cursor._ColBufferList is not getting populated properly when a query is 
executed.

I fixed this by modifying the _UpdateDesc function a bit, although I still need 
to run tests to make sure that it doesn't break anything else. Basically, I 
just moved the following line to the *end* of the function, instead of it being 
inside the if statement:

self._row_type = self.row_type_callable(self)

This forces the Row object to be created *after* the _ColBufferList is 
populated, which fixes the TypeError exception with NamedTupleRow and 
MutableNamedTupleRow.

I'll try to submit some kind of patch, but this really is just a one line 
change.

Original comment by [email protected] on 24 Sep 2014 at 7:08