Force a Validate before saving your edits

04-22-2010 07:25 PM
Status: Open
New Contributor II

Most of the great features (Topologies, Networks, Domains, etc.) that provide an extra edge to your spatial information in terms of data QA/QC can sometimes be completely overridden if the appropriate validation is not performed (i.e. FieldCalculator enables entries outside of Domains, topologically invalid features can still be saved if not checked against an existing Topology, etc.).
It would be appropriate to give the Data Owner of such objects the option to enforce such a "Validate" upon a user before he/she is allowed to save and exit the editing session, until all edits are conform.

Tags (1)
I strongly agree!  In my agency, we have corporate data layers that are updated by multiple editors.  In the "old days" we used coverages and AMLs we could make sure that data passed all "hard" editchecks before being posted to the production database.

Now, I am working on a project to replace one of those legacy edit applications.  In the data model, we are using domains, subtypes, and topology rules to help enforce data integrity from the get-go and help keep bad data from ever getting in.  (These tools are fantastic and keep getting better all the time).  However, these tools do not help us guarantee data integrity.  So, we are having to duplicate all these checks using Python scripts or other methods.  Some of the checks are "soft" checks which can exist and the data will still be allowed to be posted.  But if the data fails a "hard" check, the editor must fix the problem before they can post it.  The only other way to do it would be to use a human "gate keeper" who runs the checks before posting but we do not have the resources for this.

I understand the good side to not enforcing topology (and the flexibility it provides to users) but since we cannot use the software to enforce the rules we want it to, we have to do a lot more work developing the same checks and not allowing data to enter the system until it is good enough to pass QAQC.
Something to consider is using Data Reviewer to configure checks stored in a batch job that can be run regulary.  And if you use Workflow manager, you can create a workflow where the editors will have to execute the batch job after their edits.  If there are any errors that have a high priorty (defined in the check), those would have to be addressed before moving on to the next step.