Truncated values when accessing decimal fields when using the GeoDjango DataSource

[khoobks]

Hi Everybody,

I noticed what appeared to be a truncation issue today while attempting to retrieve decimal fields from a Map Info TAB file using the GeoDjango datasource. 

I'd like to find out if 

1) This is replicable for other people (I have replicated it successfully on two machines)

2) If this is a known issue and

3) If there are any workarounds besides falling back to 'normal' OGR.

I have attached a sample TAB file and a python file to test the problem. The output on my machine is as follows,

Python 2.7.2+ (default, Oct  4 2011, 20:06:09) 

[GCC 4.6.1] on linux2

Type "help", "copyright", "credits" or "license" for more information.

(InteractiveConsole)

>>> filepath = '/tmp/test_datatypes/test_datatypes.TAB'

>>> 

>>> # Using Django

>>> import django

>>> print django.VERSION

(1, 4, 1, 'final', 0)

>>> from django.contrib.gis.gdal import DataSource

>>> ds = DataSource(filepath)

>>> lyr = ds[0]

>>> for feat in lyr:

...     print '----------'

...     print 'Feature ID: ', feat.fid

...     for field_name in feat.fields:

...         print 'Field Name: ', field_name

...         print 'Field Data:', type(feat.get(field_name)), feat.get(field_name)

...         print

... 

----------

Feature ID:  1

Field Name:  char

Field Data: <type 'str'> test

Field Name:  deci

Field Data: <type 'float'> 1.234

Field Name:  int

Field Data: <type 'int'> 5

Field Name:  date

Field Data: <type 'datetime.date'> 2012-08-24

Field Name:  float

Field Data: <type 'int'> 987

----------

Feature ID:  2

Field Name:  char

Field Data: <type 'str'> abc

Field Name:  deci

Field Data: <type 'float'> 987.6

Field Name:  int

Field Data: <type 'int'> 3

Field Name:  date

Field Data: <type 'datetime.date'> 2012-08-26

Field Name:  float

Field Data: <type 'int'> 12345

>>> # Using OGR

>>> import ogr

>>> ds = ogr.Open(filepath)

>>> lyr = ds[0]

>>> for feat in lyr:

...     feat.items()

... 

{'char': 'test', 'int': 5, 'float': 987.65, 'date': '2012/08/24', 'deci': 1.234}

{'char': 'abc', 'int': 3, 'float': 12345.6789, 'date': '2012/08/26', 'deci': 987.6}

>>> 

The problem is that for the field named "float" the expected value is 987.65 for the first feature and 12345.6789 for the second field. Using geodjango however I am getting 987 and 12345 respectively. This is of course a truncation of some sort making the number into an int. 

Has anyone else experienced this problem? Is this a known issue and is the appropriate work around simply to fall back to 'normal' OGR?

On a separate note purely for my own edification, does anyone know why the design decision was made to wrap OGR and GDAL and not simply reuse the python bindings that they provide?

Thanks in advance for any help that you can provide.

Ben