read_file on an empty layer in a gdb results in a KeyError

I’m not sure if this is expected behavior, but I was trying to read a specific layer in a geodatabase which happened to be empty (has no rows, but has columns) using geopandas.read_file() and it resulted in a KeyError. The same gdb and the layer within the gdb can be obtained from here: https://coast.noaa.gov/htdata/Inundation/SLR/SLRdata/Pacific/HI_Lanai_slr_data_dist.zip

Full stack trace below:

gdf = geopandas.read_file('../noaa/sea_level_rise/hawaii/HI_Lanai_slr_data_dist/HI_Lanai_slr_final_dist.gdb', layer='HI_Lanai_low_2ft')

---------------------------------------------------------------------------
KeyError                                  Traceback (most recent call last)
<ipython-input-7-513ea4fddcc2> in <module>()
      1 gdf = geopandas.read_file('../noaa/sea_level_rise/hawaii/HI_Lanai_slr_data_dist/HI_Lanai_slr_final_dist.gdb',
----> 2                          layer='HI_Lanai_low_2ft')

/usr/local/lib/python3.6/site-packages/geopandas/io/file.py in read_file(filename, **kwargs)
     28         # re-order with column order from metadata, with geometry last
     29         columns = list(f.meta["schema"]["properties"]) + ["geometry"]
---> 30         gdf = gdf[columns]
     31 
     32     return gdf

/usr/local/lib/python3.6/site-packages/geopandas/geodataframe.py in __getitem__(self, key)
    396         GeoDataFrame.
    397         """
--> 398         result = super(GeoDataFrame, self).__getitem__(key)
    399         geo_col = self._geometry_column_name
    400         if isinstance(key, string_types) and key == geo_col:

/usr/local/lib/python3.6/site-packages/pandas/core/frame.py in __getitem__(self, key)
   2051         if isinstance(key, (Series, np.ndarray, Index, list)):
   2052             # either boolean or fancy integer index
-> 2053             return self._getitem_array(key)
   2054         elif isinstance(key, DataFrame):
   2055             return self._getitem_frame(key)

/usr/local/lib/python3.6/site-packages/pandas/core/frame.py in _getitem_array(self, key)
   2095             return self.take(indexer, axis=0, convert=False)
   2096         else:
-> 2097             indexer = self.ix._convert_to_indexer(key, axis=1)
   2098             return self.take(indexer, axis=1, convert=True)
   2099 

/usr/local/lib/python3.6/site-packages/pandas/core/indexing.py in _convert_to_indexer(self, obj, axis, is_setter)
   1228                 mask = check == -1
   1229                 if mask.any():
-> 1230                     raise KeyError('%s not in index' % objarr[mask])
   1231 
   1232                 return _values_from_object(indexer)

KeyError: "['Id' 'grid_code' 'Shape_Length' 'Shape_Area' 'geometry'] not in index"

Are there any work arounds? or do I have to try-except the KeyError which seems a little broad.

[kuanb]

This is hacky, and does not solve the problem, but might at least improve your KeyError try .. catch logic:

try:
    gdf = gpd.read_file(filepath)
except KeyError as exception:
    with fiona.open(dir_data, driver="OpenFileGDB") as src:
        if _ensure_that_there_are_zero_rows(src):
            pass
        else:
            raise exception

With this method, you would at least be able to try a second attempt at reading the GDB and double checking that there indeed are no rows.

[jorisvandenbossche]

We recently fixed a bug for reading empty files (https://github.com/geopandas/geopandas/issues/649, PR https://github.com/geopandas/geopandas/pull/653), I suppose this will be the same.