Turn child replica into parent?

[KarenFolger]

What's the best way to go about populating a replica with data and then turning that version into the parent for 1-way replication back to what was the parent?  We both have 10.1 SDE and want to populate a database with selected data from the parent.  Then, we want to make the child the 1-way parent back to what was the parent.  If there is a simpler method to maintaining the original globalID's that would be helpful, but this seems to be the only way.

[CherylCleghorn]

Karen

It is not clear if you already have a replica or if you are now creating a replica. If you are now creating a replica then you can create a 1-way child-to-parent replica. If you already have a 1-way parent-to-child replica then you cannot switch the parent-child roles. Your recourse would be to synchronize all your changes (parent-to-child), unregister the existing replica and then recreate it as 1-way child-to-parent. You can then make your edits in the child and synchronize with the parent.

See Replication Types

Cheryl

[KarenFolger]

Karen

It is not clear if you already have a replica or if you are now creating a replica. If you are now creating a replica then you can create a 1-way child-to-parent replica. If you already have a replica then you cannot switch the parent-child roles. Your recourse would be to synchronize all your changes, unregister the existing replica and then recreate it as 1-way child-to-parent.

See Replication Types

Cheryl

version 10.1 - we have not yet created any replicas.  We have a situation where a regional office has the master data but we want to send parts of the data to field offices and then have the field office be responsible for their piece of the data and replicate it back up to the regional office.  So part of the "problem" for us is, how do we maintain the global id's from the master dataset and then turn it around so that the field is the master (of only their piece of the data) but populates the larger whole at a regional level?

[CherylCleghorn · Accepted solution]

In that case Karen you can create a 1-way child-to-parent replica and define the extent that you want for the field. The child however must be an sde geodatabase. Note that if within that extent, edits are made by the parent then those edits will be overwritten when the field synchronizes edits back.

To preserve globalids you can have the data copied during the replica creation process or you can use one of the methods noted under globalid in the following reference to copy the data, and then use the Register Existing Data option to create the replica.

ArcGIS field data types


extract:

"Copying and pasting, data extraction, and XML workspace exporting and importing preserve global ID values in the output geodatabase. Other data exporting and importing methods do not preserve these values."

Cheryl

[KarenFolger]

Thanks - good info. 

I think our biggest challenge at this point is trying to figure out the best workflow for getting the initial set of selected area data to the field.  We could create the field data as a child of the parent and then break the replica and re-create in the other direction.  That seems simplest to me but I might be missing somethings.

We're both in SDE but the regional offices want to ensure that the field offices only get control of the data that is for them in the field.  So the initial population to the field is the critical piece.

[CherylCleghorn]

Karen,

No need to break and re-create the replica. When you are creating the 1-way replica, choose the child-to-parent option, as per step 6 here:
Creating a one-way or two-way replica

This sets the child replica as the one that is able to edit and sync changes back to the parent (which is what you want). Don't be confused by the term parent/child once the replica is created; think instead in terms of which replica can edit and sync changes back. I like to think of the parent replica geodatabase as the geodatabase from which the data originated.

Regards
Cheryl