Whats the best way to reconcile AGOL edits with a geodatabase in Pro?

Question

I'm interested to hear how people are best utilising AGOL and Pro to reconcile any edits that have been made in AGOL, from a feature class that was previously published from a GDB in Pro.

One of the simplest approaches might be to simply copy and overwrite the original version. But if you had published the feature service with a sub-set of fields switched on, you'd then lose these in the overwrite.

Another approach might be to append the data with field mapping enabled (if you have a sub-set of fields).

Or if you're just updating attributes, set up a python script to join the data, copy the data over and then remove the join.

Does anyone else have a good suggestion? For a workflow I'm currenty working on, I have published data from a GDB to AGOL, and surveyors are "marking up" by creating new features AND editing attributes of existing features. So I'm steering towards Append.

David_Brooks · Answer

@jcarlson yeah that's a good approach for managing repeated edits.

Re: your first question. There aren't direct edits happening of course, but I'm looking at the best way to pull down the deltas from the hosted service into the original GDB FC.

Basically, I have 2 hosted feature services. One is published empty for surveyors to digitise with, and I will be appending the new features back into the GDB FC. The other is published with a bunch of features in that will have attributes edited, and possibly geometries. And I want to find the best way to reconcile those edits in AGOL with original FC in the GDB. Copy is not an option because the schemas don't match, so the options are JOIN for attribute only, or some kind of append and replace script.

jcarlson · Answer

If a feature class is published as an editable service, there really shouldn't be edits happening back in the source GDB. What exactly is in the GDB that isn't in the published service?

I will say, overwriting is usually not the best approach, especially if there may be schema changes or hosted view layers involved.

For us, we do have a couple of hosted layers that need to update from a separate source, though the hosted layers are not edited. We do this by using editor tracking and a shared GUID field so that features from each can be uniquely identified and linked. Then we use the ArcGIS Python API to query out changes based on the last_edited_date field and apply those features as edits to the hosted layer.

It's a bit much to explain in text, but the end result is a process that selectively identifies edits and applies them in place. For instance, we have about 60k parcels, but if I only edit 3 of them today, the script will only apply edits to three features in the associated hosted layer.