In a versioned GDB, if I create new topology on a feature dataset in my edit version, validate and correct errors, then remove the topology layer from the feature dataset, would doing this affect the default during rec and post?

[Anonymous User]

Working with geometric networks in a multi-user environment and would like to use topology rules without un-versioning the entire project.  The project is already well under way, and un-versioning in order to add new topology rules is out of the question at this point. 

My work around is to add topology rules on my version, validate them, and then remove the topology from the feature dataset prior to rec and post.  Is this too risky?  Should I attempt it in a test GDB or just scrap the idea?  

The reason I am suggesting this work around is that when I was recently learning how to build topology rules, I used GDB backups of finished data to practice and found errors in the data using my topology rules.  This proves the value of topology rules, but un-versioning is out of the question.  

Thank you in advance.

[DanaNolan]

I haven't tried this. I added topology to an enterprise dataset, gave up trying to get it to validate in a version, then did topology validation in a file geodatabase instead and replaced all the records. Your workaround sounds OK. Any changes you make to fix topology errors are edits, so they should rec and post fine. There are two issues though.

If your data has a lot of errors, rules, or changes a lot, you could be re-marking a lot of exceptions. You can export your topology errors, so you can see what you marked the last time, but saved topology data does not have all the functionality of a feature class such as joins, which you could use to help determine which errors are new. 

The other issue is documentation/metadata. Make sure your rules are saved out to a permanent template database and documented, so that everyone knows they exist even though they won't show in ArcCatalog for your enterprise FDS.

[George_Thompson]

My question is why do you want to remove topology before rec and post?

I would not see it having any impact on the underlying data, but as Dana Nolan‌ mentions you may have to re-mark a lot of exceptions.