In one of our enterprise geodatabases, a single feature class within a feature dataset fails to draw and tosses the above mentioned error. (10.3.1) There are a few older Geonet postings and even a Tech Support article that indicate the problem is between the features of the feature class exceeding some boundary set by the feature dataset.
I exported the problem child feature class from the e-gdb to a local file geodatabase (10.5.1). Draws just fine. Then I created a feature dataset with the same spatial coordinate system as the the original enterprise feature dataset; If I try to copy and past the exported feature class into my new feature dataset, it fails with the famous "Spatial references do not match" error. Never have understood that one, but whatever. Of course, I can import the local feature class into the local feature dataset, and everything is happy.
Just for fun, I copied the enterprise geodatabase feature dataset and pasted it into my local fgdb. The former problem child is happy as a clam; draws all day long. All the feature classes within that feature dataset are happy.
So what's going on here? How does one feature class in a feature dataset all of the sudden decide it has a conflict with it's feature data set? I think we can delete and then import a new copy of the problem child feature class, but that's no fun; we still don't know what happened.....