I have services running for our town that represent town parcels for our interactive mapping website. Internally and for private use this parcel layer (let's call is Parcel 1) contains roll number, owner name, address, and various other fields. We have another layer that is basically just a copy of Parcel 1 but is for public use. This layer contains all the same fields except for the owner information (Parcel 2). What I have done is basically copied Parcel 1 and deleted the unnecessary fields for Parcel 2.

When changes are made to Parcel 1, I delete Parcel 2, export/copy Parcel 1 and rename it to Parcel 2 (or simple edit the geometry and information in Parcel 2 to reflect changes in Parcel 1...essentially making the same edits twice.)

What I want to know is, is this the most efficient workflow?

What I want is for when changes are made to Parcel 1 for Parcel 2 to automatically update it's geometry and fields at the same time.

Is this possible since there is different schema?

I have replicas running on these feature classes so add/deleting a feature class proves to be quite a pain.

The comments below to create separate services with fields hidden seems to work fine. The answer for creating an additional replica for Parcels also seems feasible. The problem I'm encountering now is that the parcel data is bad quality. We receive a lot of our information from a tax database and another source and the road names are not correct or contain different alias. What I did was create a a copy of the parcels and renamed the roads so that a query/search by the user would yield a result. As a work around. I'm uncertain how to proceed. A replica with the street name field I created (different from Parcel1) could feasibly work so that the original street name isn't synced. However, new parcels that are added would then not be synchronized correctly.

Any thoughts on this issue?

  • 1
    My gut reaction would be to use Parcel 1 for both purposes, but restrict the access of the public website so it isn't able to see the restricted fields (parcel owner). This may not be feasible, of course :) The question is interesting apart from that, however, and I'm interested to see answers/ideas.
    – Erica
    Jun 17, 2014 at 13:25
  • I agree with Erica, have two different services from the same feature class. For the public version just turn off the owner field in the Layer Properties>Field tab, then publish.
    – artwork21
    Jun 17, 2014 at 13:29
  • I'll give this a try - I believe I tried that initially but there were instances where the field would still show up. We are using Geocortex. Thanks!
    – GISHuman
    Jun 17, 2014 at 13:31

1 Answer 1


Well, for tags you've marked ArcGIS Server and you discuss replicas, so I'm assuming you have access to SDE and am assuming in this answer you are using an SDE (versioned) database for storing and editing at least the Parcel 1 layer. Before I continue, let me say that, if possible, just publishing two different services with different fields turned on seems rather ideal, but, if that is not a good option for you and you need a separate Parcel 2 database/feature class, you can try the following. What you could do is make a fresh copy of parcel 2 feature class from parcel 1. But do this by copying the entire parcel 1 feature class rather than appending in features or anything like that as the goal of this should be to preserve the same global id for each feature in both feature classes. Then, in the new parcel 2 layer, delete the un-wanted fields. Then, just set up a replica, preferably one way from parcel 1 to parcel 2. If you do that, any time you make edits in parcel 1, you'd just have to synch the replica and parcel 2 would have those edits (either geometry or attribute edits) but would still not have the restricted info as there would be no field to store it in. You can replicate across different schemas, the limitation is it will only replicate fields that have the same name and field type/length on both ends of the replica. The problem with making that a 2 way replica, is that if you make the edit in parcel 2, depending on how you synchronize things and such, I don't know what would happen to the extra fields for the changed features in parcel 1 when synchronized, as parcel 2 wouldn't have that extra data to send to parcel 1.

Anyway, I know that's a long, wordy answer and may be as clear as mud, but, if you need clarification or help, just let me know.

EDIT: Now, I think you second part of the question would honestly be better as a separate question, but, I will give at least an initial thought on that here as well. My initial reaction would be, if you are worried about editing the original road names in your original file, but still want the corrected data available, why not either add an additional field for "newroadname" or such as that, or, you could use a joined or related table based on some unique ID join to create a list of road name aliases. So, you could either have an original road listed as Smith St, but you could have a related table (set up relationship class in database) that lists alternate road names for that same record as Smith Dr, George Smith Dr, and/or County Road 123. Now, I'm a bit less clear on exactly what you were asking for on this part of your question, so I don't know if that actually answers the question at all, but, based on what I think you were asking I have given the answer. Again, I suggest either further clarifying the edited part of the question, or, ideally, posting that second part of the question as a whole new question. I think you'd get potentially a better variety and/or quality of answers.

  • That actually makes a lot of sense. We have quite a few replicas running already. I have another issue though, which I will be adding to my question via edit!
    – GISHuman
    Jun 17, 2014 at 14:07
  • @GISKid See edited answer.
    – John
    Jun 17, 2014 at 14:50
  • Thanks for all your help John. A second question might be in order, I am still interested in other alternatives for the first part as well, though. I think part of the issue is creating workflows and I don't believe they would work on related tables.
    – GISHuman
    Jun 17, 2014 at 15:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.