Publishing dataframe with missing geometry via the GeoAccessor causes inconsistent behavior

Describe the bug

When using GeoAccessor.to_featurelayer, I get a FeatureCollection returned and the dataframe is not published to a feature layer (ArcGIS Online) or the process throws error code 500 (Portal 10.7.1).

To Reproduce Steps to reproduce the behavior:

from arcgis.features import GeoAccessor

lyr = df.spatial.to_featurelayer('dwelling events', gis=agol)
lyr
# returns: <FeatureCollection>

type(lyr)
# returns: arcgis.features.feature.FeatureCollection

s = agol.content.search('dwelling events'); s
# returns: [ ]

Note: when passing a GIS object connected to a 10.7.1 Portal as the gis parameter, throws 500 internal server error:

---------------------------------------------------------------------------
Exception                                 Traceback (most recent call last)
<ipython-input-51-f6f9e09801c1> in <module>
      1 from arcgis.features import GeoAccessor
      2 
----> 3 lyr = final_df.spatial.to_featurelayer("dwelling_events", gis=gis)
      4 lyr

/usr/local/anaconda3/envs/spark/lib/python3.7/site-packages/arcgis/features/geo/_accessor.py in to_featurelayer(self, title, gis, tags, folder)
   2132                 raise ValueError("GIS object must be provided")
   2133         content = gis.content
-> 2134         return content.import_data(self._data, folder=folder, title=title, tags=tags)
   2135     # ----------------------------------------------------------------------
   2136     @staticmethod

/usr/local/anaconda3/envs/spark/lib/python3.7/site-packages/arcgis/gis/__init__.py in import_data(self, df, address_fields, folder, item_id, **kwargs)
   4935             if item_id:
   4936                 postdata['itemIdToCreate'] = item_id
-> 4937             res = self._portal.con.post(path, postdata)#, use_ordered_dict=True) - OrderedDict >36< PropertyMap
   4938 
   4939             fc = FeatureCollection(res['featureCollection']['layers'][0])

/usr/local/anaconda3/envs/spark/lib/python3.7/site-packages/arcgis/gis/_impl/_con/_connection.py in post(self, path, params, files, **kwargs)
    708                                      file_name=file_name,
    709                                      try_json=try_json,
--> 710                                      force_bytes=kwargs.pop('force_bytes', False))
    711     #----------------------------------------------------------------------
    712     def put(self, url, params=None, files=None, **kwargs):

/usr/local/anaconda3/envs/spark/lib/python3.7/site-packages/arcgis/gis/_impl/_con/_connection.py in _handle_response(self, resp, file_name, out_path, try_json, force_bytes)
    503             if 'error' in data:
    504                 errorcode = data['error']['code'] if 'code' in data['error'] else 0
--> 505                 self._handle_json_error(data['error'], errorcode)
    506             return data
    507         else:

/usr/local/anaconda3/envs/spark/lib/python3.7/site-packages/arcgis/gis/_impl/_con/_connection.py in _handle_json_error(self, error, errorcode)
    521 
    522         errormessage = errormessage + "\n(Error Code: " + str(errorcode) +")"
--> 523         raise Exception(errormessage)
    524     #----------------------------------------------------------------------
    525     def post(self,

Exception: Internal Server Error
(Error Code: 500)

Expected behavior

1. Return type should be: arcgis.features.layer.FeatureLayer
2. Feature layer should be published to org/portal, respecting given parameters (title, folder, tags, gis).
3. Publishing behavior/responses should (ideally) be consistent across AGOL/Portal
4. If there is a problem with the DataFrame or publishing process, publishing/conversion should fail noisily

From the docs:

to_featurelayer
- publishes a spatial dataframe to a Feature Layer

Argument | Description
-----------------------
title - Required string. The name of the service
gis - Optional GIS. The GIS connection object
tags - Optional list of strings. A comma seperated list of descriptive words for the service.
folder - Optional string. Name of the folder where the featurelayer item and imported data would be stored.

returns FeatureLayer

Platform (please complete the following information):

• OS: macOS Catalina
• Browser [e.g. chrome, safari]: Chrome
• Python API Version: 1.8.2

Additional context Potentially related to #682

The DataFrame is the output of a notebook almost identical to this one for finding dwell locations put out by the tracker team, but executed on some of the natural resources team’s tracks and polygons

Zipped pickle of the DataFrame: df.zip

Conda env.yml (as .txt): env.txt

[achapkowski]

Does gis.content.import_data work with the SeDF?

[nanaeaubry]

Fixed in 2.0.0