I have a hosted feature service with a related table. The layer in the feature service needs to be updated every month or so. However, the data in the related table needs to remain unchanged. Is there a pythonic way to update the layer without altering the related data? I looked at creating an .sd file with Python but I don't anticipate much success based on what testing I've done. Would deleting the layer's contents and updating through the REST be feasible when I'm dealing with +70 thousand spatial objects?
If I understand the question correctly, you'll need to create a separate .mxd file with nothing more than the feature you'd like to update in it. Then, from that new .mxd file, publish a feature service. After that initial push, use this python script called update-hosted-feature-service to overwrite the hosted feature service as you schedule it to run.