QuerySet.defer does not work properly on Oracle

[Jani Tiainen]

The machinery in OracleQuery.resolve_columns does not take into account any fields that are deferred. Seeing this error depends on the order of the fields, and how they are handled by OracleQuery.convert_values -- which is probably why it's missed in the test suite. Here's an example model that demonstrates the problem:

from django.db import models

class MyModel(models.Model):
     num = models.FloatField()
     geom = models.CharField(max_length=128)

Here's how to raise it:

>>> MyModel.objects.create(num=5.0, name='Foobar')
<MyModel: MyModel object>
>>> qs = MyModel.objects.defer('num')
>>> print qs
Traceback (most recent call last):
  File "<ipython console>", line 1, in <module>
  File "C:\Python25\lib\site-packages\IPython\Prompts.py", line 552, in __call__
    manipulated_val = self.display(arg)
  File "C:\Python25\lib\site-packages\IPython\Prompts.py", line 578, in _display
    return self.shell.hooks.result_display(arg)
  File "C:\Python25\lib\site-packages\IPython\hooks.py", line 141, in __call__
    ret = cmd(*args, **kw)
  File "C:\Python25\lib\site-packages\IPython\hooks.py", line 171, in result_display
    out = pformat(arg)
  File "C:\Python25\lib\pprint.py", line 111, in pformat
    self._format(object, sio, 0, 0, {}, 0)
  File "C:\Python25\lib\pprint.py", line 129, in _format
    rep = self._repr(object, context, level - 1)
  File "C:\Python25\lib\pprint.py", line 195, in _repr
    self._depth, level)
  File "C:\Python25\lib\pprint.py", line 207, in format
    return _safe_repr(object, context, maxlevels, level)
  File "C:\Python25\lib\pprint.py", line 292, in _safe_repr
    rep = repr(object)
  File "C:\django\trunk\django\db\models\query.py", line 61, in __repr__
    data = list(self[:REPR_OUTPUT_SIZE + 1])
  File "C:\django\trunk\django\db\models\query.py", line 76, in __len__
    self._result_cache.extend(list(self._iter))
  File "C:\django\trunk\django\db\models\query.py", line 261, in iterator
    for row in self.query.results_iter():
  File "C:\django\trunk\django\db\models\sql\query.py", line 292, in results_iter
    row = self.resolve_columns(row, fields)
  File "C:\django\trunk\django\db\backends\oracle\query.py", line 54, in resolve_columns
    values.append(self.convert_values(value, field))
  File "C:\django\trunk\django\db\backends\oracle\query.py", line 74, in convert_values
    value = float(value)
ValueError: invalid literal for float(): Foobar

The problem reveals itself here because although we've deferred the FloatField, it's instance is still passed into convert_values with the value associated with the CharField.

[jbronn]

This problem only affects the Oracle spatial backend.

[jbronn]

The problem is not GeoDjango related at all, and applies to the entire Oracle backend.