I know you're disabling it mid-step, but can you test it with a duplicate feature layer that doesn't have sync enabled at all? Sounds like it is still storing previous versions or a history of previous edits somewhere, and if you're doing a full truncate/append that's a lot of "edits".
When I need to do full overwrites I either use the .update() method to replace the whole file, or use spatially enabled dataframes (SEDF) and .edit_features(adds='your SEDF'). Something like the examples below. Could try one of these as an alternative to append() .
### Update Hosted Feature Layers Examples
from arcgis.gis import GIS
from arcgis.features import FeatureLayerCollection
updated_data = r'this_is_a_path_to_file.csv' # path to your updated dataset, the file type must match whatever the service was published with
itemid = '' # the itemid for the hosted feature service you want to update in ArcGIS Online, you can get it off item page in AGO
portal_url = '' # "https://example.maps.arcgis.com"
user = '' # AGO username
pw = '' # AGO password
### Log in to ArcGIS Online account
gis = GIS(portal_url, user, pw)
un = gis.properties.user.username
print(f'\nLogged in as {un} on {portal_url} \n')
#############################################################################
### Example 1 (Overwrite directly from local file)
# If you published to AGO from a local file (csv, fgdb, etc.), this is the simplest way to overwrite
# The new data must have the same filename and file type, and FGDBs must be zipped
hostedfeature = gis.content.get(itemid)
flayer = FeatureLayerCollection.fromitem(hostedfeature)
flayer.manager.overwrite(updated_data)
print("Hosted feature layer updated with the latest and greatest")
# That is the barebones method, but you can update lots of other stuff too, for example the description, title, etc.
# https://developers.arcgis.com/rest/users-groups-and-items/common-parameters/
item_properties={
"description":"Here is my description, isn't it great",
"title":"Great Item Title"
}
hostedfeature.update(item_properties)
# Or the thumbnail
thumb = r'path_to_some_picture.jpg' # can also use a image url
hostedfeature.update(thumbnail=thumb)
# Or you can disable and enable settings, like Sync
syncoff = {"capabilities": "Query, Extract"}
hostedfeature.manager.update_definition(syncoff)
syncon = {"capabilities": "Query, Extract, Sync"}
hostedfeature.manager.update_definition(syncon)
# https://developers.arcgis.com/python/latest/samples/overwriting-feature-layers/#:~:text=updated_capitals_csv%27%2C%20csv_file_name))-,Overwrite%20the%20feature%20layer,-Let%20us%20overwrite
#############################################################################
### Example 2 (Go the edit route with a full truncate and append)
# You can do this either from a regular pandas dataframe if it's a table, a spatially enabled dataframe if there is geometry, or from a featureset
# If you are adding more than 1000 features you want to chunk it out
import pandas as pd
import os
hostedfeature = gis.content.get(itemid).layers[0] # if you have a hosted table you can use .tables[0] instead
hostedfeature.manager.truncate()
local_featureclass = os.path.join(r"path to a file geodatabase .gdb", "MyFeatureClass")
sedf = pd.DataFrame.spatial.from_featureclass(local_featureclass)
n = 1000
for i in range(0, len(sedf), n):
chunk = sedf[i:i+n]
result = hostedfeature.edit_features(adds=chunk)
print(f'added items {i+1}-{i+n}...')
-Aaron
