I have a feature class in versioned geodatabase that contains a status field that flags the record as "Proposed" or "Completed". The data users want this published as two separate feature classes based on the status field.
I am looking at using a 1-way replica with a definition query filter (based on the status field) to separate the data into the two publication feature classes (in file geodatabases, but possibly also non-versioned SDE feature classes).
When a user makes an edit to change a record from Proposed to Completed, the subsequent synchronization adds that record to the publication Completed feature class, but it does not remove that record from the publication Proposed feature class - it remains there (with it's new status of Completed)... so now the same record is in both publication feature classes.
My initial thought is to sweep through the publication feature classes after synchronizing and delete that record where the status was updated and no longer meets the criteria of the definition query (in the example above, delete the now Completed record that exists in the publication Proposed feature class).
Does this sound like a valid way to accomplish this request by the data users, or am I possibly going to break the replica? Early tests seem OK but I've only done a few syncs/deletes.