Spatially Enabled Data Frame not working with pandas datetime

Describe the bug Attempting sdf = pd.DataFrame.spatial.from_layer() returns a warning “FutureWarning: The pandas.datetime class is deprecated and will be removed from pandas in a future version. Import from datetime instead.” and gives me “Exception: Could not load the dataset: dtype ‘<class ‘pandas.__DatetimeSub’>’ not understood”

Are there any workarounds to downloading AGOL data with Python API? sdf = SpatialDataFrame.from_layer() is the deprecated version which also returns an error.

I’ve been running into an issue much like what’s happened here (code is essentially the same): https://community.esri.com/t5/arcgis-api-for-python-questions/arcgis-api-for-python-spatially-enabled-data-frame-broken-with/td-p/865661

To Reproduce Steps to reproduce the behavior:

from arcgis.gis import *
from arcgis.features import GeoAccessor, GeoSeriesAccessor

gis = GIS("https://arcgis.com", ***, ***)
item = gis.content.get('The item id')
QB = item.layers[0]

# create a Spatially Enabled DataFrame object
sdf = pd.DataFrame.spatial.from_layer(QB)
sdf.head()

error:

TypeError                                 Traceback (most recent call last)
~\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone1\lib\site-packages\arcgis\features\geo\_accessor.py in from_layer(layer)
   1966             from arcgis.features.geo._io.serviceops import from_layer
-> 1967             return from_layer(layer=layer)
   1968         except ImportError:

~\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone1\lib\site-packages\arcgis\features\geo\_io\serviceops.py in from_layer(layer, query)
    122     else:
--> 123         return layer.query(where=query, as_df=True)#.sdf
    124     return res

~\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone1\lib\site-packages\arcgis\features\layer.py in query(self, where, out_fields, time_filter, geometry_filter, return_geometry, return_count_only, return_ids_only, return_distinct_values, return_extent_only, group_by_fields_for_statistics, statistic_filter, result_offset, result_record_count, object_ids, distance, units, max_allowable_offset, out_sr, geometry_precision, gdb_version, order_by_fields, out_statistics, return_z, return_m, multipatch_option, quantization_parameters, return_centroid, return_all_records, result_type, historic_moment, sql_format, return_true_curves, return_exceeded_limit_features, as_df, **kwargs)
    735                 import pandas as pd
--> 736                 df = self._query_df(url, params)
    737                 dt_fields = [fld['name'] for fld in self.properties.fields \

~\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone1\lib\site-packages\arcgis\features\layer.py in _query_df(self, url, params)
   1478                     df.loc[q, fld['name']] = 0
-> 1479         df = df.astype(dtypes, False)
   1480         if 'geometryType' in featureset_dict:

~\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone1\lib\site-packages\pandas\core\generic.py in astype(self, dtype, copy, errors)
   5530                     results.append(
-> 5531                         col.astype(dtype=dtype[col_name], copy=copy, errors=errors)
   5532                     )

~\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone1\lib\site-packages\pandas\core\generic.py in astype(self, dtype, copy, errors)
   5545             # else, only a single dtype is given
-> 5546             new_data = self._mgr.astype(dtype=dtype, copy=copy, errors=errors,)
   5547             return self._constructor(new_data).__finalize__(self, method="astype")

~\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone1\lib\site-packages\pandas\core\internals\managers.py in astype(self, dtype, copy, errors)
    594     ) -> "BlockManager":
--> 595         return self.apply("astype", dtype=dtype, copy=copy, errors=errors)
    596 

~\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone1\lib\site-packages\pandas\core\internals\managers.py in apply(self, f, align_keys, **kwargs)
    405             else:
--> 406                 applied = getattr(b, f)(**kwargs)
    407             result_blocks = _extend_blocks(applied, result_blocks)

~\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone1\lib\site-packages\pandas\core\internals\blocks.py in astype(self, dtype, copy, errors)
    547         if dtype is not None:
--> 548             dtype = pandas_dtype(dtype)
    549 

~\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone1\lib\site-packages\pandas\core\dtypes\common.py in pandas_dtype(dtype)
   1777     elif npdtype.kind == "O":
-> 1778         raise TypeError(f"dtype '{dtype}' not understood")
   1779 

TypeError: dtype '<class 'pandas.__DatetimeSub'>' not understood

During handling of the above exception, another exception occurred:

Exception                                 Traceback (most recent call last)
<ipython-input-48-b148856d5a67> in <module>
      1 # Convert to Spatial Data Frame to make a local data source for processing
----> 2 sdf = pd.DataFrame.spatial.from_layer(QB)
      3 #sdf.to_featureclass("QB_Output")
      4 
      5 #QBData.download(save_path=r"C:\Data\GIS_Projects_Local\ACHRIS\AGOL_to_QB") #downloads as 'file', not readable

~\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone1\lib\site-packages\arcgis\features\geo\_accessor.py in from_layer(layer)
   1971             raise Exception("Malformed response from server, could not load the dataset: %s" % str(je))
   1972         except Exception as e:
-> 1973             raise Exception("Could not load the dataset: %s" % str(e))
   1974     #----------------------------------------------------------------------
   1975     @staticmethod

Exception: Could not load the dataset: dtype '<class 'pandas.__DatetimeSub'>' not understood

Platform (please complete the following information):

• Browser Chrome
• Python API Version 1.5.3

[AtmaMani]

If you are using Pro, you may have to clone your environment and then update the API. Start the ‘Pro Python command prompt’ with elevated privileges (as administrator) and run

conda create --name myclone --clone arcgispro-py3

then activate it

conda activate myclone

then install Python API. You may have to unpin as shown below:

conda upgrade --no-pin arcgis

[mw1m]

Thanks for your help @priyankatuteja and @AtmaMani Updating using the conda upgrade --no-pin arcgis ended up doing the trick!