AttributeError when converting DataFrame to GeoDataFrame with Polygon

Question

I imported a shapefile in the form of a pandas.DataFrame and ran operations on it. But I can't convert the DataFrame to a GeoDataFrame. I tried with gpd.GeoDataFrame, and with wkt.loads. I think it's because of the format of my geometry: they're Polygon Z:

0 POLYGON Z((653467.7 6840811 81.5 ...))

Here is my code when I try to create a geometry column before the conversion to GeoDataFrame

bat['BAT_COORD']=bat['BAT_COORD'].apply(wkt.loads)

and the error :

Traceback (most recent call last):
 File "cleanshp.py", line 34, in <module>
 bat['BAT_COORD']=bat['BAT_COORD'].apply(wkt.loads)
 File "/home/windmyroof/.local/lib/python3.7/site-packages/pandas/core/series.py", line 4200, in apply
 mapped = lib.map_infer(values, f, convert=convert_dtype)
 File "pandas/_libs/lib.pyx", line 2388, in pandas._libs.lib.map_infer
 File "/home/windmyroof/.local/lib/python3.7/site-packages/shapely/wkt.py", line 10, in loads
 return geos.WKTReader(geos.lgeos).read(data)
 File "/home/windmyroof/.local/lib/python3.7/site-packages/shapely/geos.py", line 279, in read
 raise TypeError("Only str is accepted.")
TypeError: Only str is accepted.

The code with the shapefile import as a geodataframe (seems good), and an attempt of export shapefile (that fails) :

import geopandas as gpd
import pandas as pd
from shapely.ops import unary_union
from shapely.geometry import Point, Polygon
from shapely import wkt
def import_shapefile(path):
 # import shapefile using geopandas
 sjer_plot_locations = gpd.read_file(path)
 # create right dataset for geometry reconstruction
 columns = ['BAT_ID', 'BAT_COORD', 'BAT_HAUTEUR', 'Z_SOL']
 bat = gpd.GeoDataFrame(index=range(len(sjer_plot_locations.geometry[:])), columns=columns)
 for i in range(len(sjer_plot_locations.geometry[:])):
 bat.at[i, 'BAT_ID'] = sjer_plot_locations.ID[i]
 bat.at[i, 'BAT_HAUTEUR'] = sjer_plot_locations.HAUTEUR[i]
 bat.at[i, 'Z_SOL'] = sjer_plot_locations.Z_MIN_SOL[i]
 bat.at[i, 'BAT_COORD'] = sjer_plot_locations.geometry[i]
 return bat, sjer_plot_locations
bat, sjer = import_shapefile('/home/windmyroof/salome_shp/bats.shp')
n_bats = [i for i in range(len(bat.index))]
# cleaning geometry
# u = unary_union([bat.at[i, 'BAT_COORD'] for i in range(len(bat.index)) if i in n_bats])
# u = u.simplify(0.5)
bat.to_file('/home/windmyroof/salome_shp/bats_cleaned.shp')

And the error :

Traceback (most recent call last):
 File "cleanshp.py", line 33, in <module>
 bat.to_file('/home/windmyroof/salome_shp/bats_cleaned.shp')
 File "/home/windmyroof/.local/lib/python3.7/site-packages/geopandas/geodataframe.py", line 746, in to_file
 _to_file(self, filename, driver, schema, index, **kwargs)
 File "/home/windmyroof/.local/lib/python3.7/site-packages/geopandas/io/file.py", line 239, in _to_file
 schema = infer_schema(df)
 File "/home/windmyroof/.local/lib/python3.7/site-packages/geopandas/io/file.py", line 299, in infer_schema
 geom_types = _geometry_types(df)
 File "/home/windmyroof/.local/lib/python3.7/site-packages/geopandas/io/file.py", line 316, in _geometry_types
 geom_types_2D = df[~df.geometry.has_z].geometry.geom_type.unique()
 File "/home/windmyroof/.local/lib/python3.7/site-packages/pandas/core/generic.py", line 5130, in __getattr__
 return object.__getattribute__(self, name)
 File "/home/windmyroof/.local/lib/python3.7/site-packages/geopandas/geodataframe.py", line 173, in _get_geometry
 " column '%s'.)" % self._geometry_column_name
AttributeError: No geometry data set yet (expected in column 'geometry'.)

I fixed the problem using the comments. It was necessary to rename the column "BAT_COORD" to 'geometry'. Indeed importing directly in gdp and not pd works very well.

• Bera: Why arent you reading the shape into a geodataframe? gpd.read_file("path/to/shapefile.shp"). Might be easier than pandas -> geopandas
• benoitWMR: thanks for your answer, I tried and it works, but I always have the problem : I want to export the shapefile on which I made operation to a new shapefile. I want to use the gpd.to_file function for that. This function requires the argument "geometry" which I can't extract.
• mwil.me: It seems from the output string that the geometry is already a Shapely geometry object. As the error message says, loads expects a string object. I think you just setting the geometry column should be enough to create a GeoDataFrame, you don't need to convert something. Can you add the code how you tried to create the GeoDataFrame?
• benoitWMR: Yes I edited my question in which i added the code
• PolyGeo: Please cut and paste your answer from the area reserved for questions into the area reserved for answers. It is fine and encouraged to self-answer.

Answer 1

Import directly to geopandas is definitely the right move. The only step you are missing is setting which column contains the geometry! Try add this line before writing to file:

bat.set_geometry('BAT_COORD', inplace=True)