I am attempting to build a geodatabase full of features and relationship classes given to me through an XML schema. The XML is set to a UTM projection that's designed for a different continent than data, and gives us errors every time we try to load our data into it. I am trying to figure out a way to project or reproject all the feature classes into a State Plane projection so that they don't loose all of their relationship classes and won't have to be rebuilt manually.
I went into the original XML, found where it was referencing the spatial reference information and deleted the existing information. I then replaced it with spatial reference information from an XML I exported from data that was in the correct State Plane projection. I rebuilt the geodatabase from the edited XML and went through and right-clicked all the features in ArcCatalog and they still had the original UTM projection selected. Just because you have the option in that menu, I figured I'd try setting all the features to the desired projection through that menu. I have never reset projections in this manner before, I have always used the respective geoprocessing tools to do so. Once I changed all the projections in this manner I brought them into ArcMap and they all drew repeatedly in the proper place with no errors.
To do a second test, I took my original, unedited XML and built a third geodatabase with all features in the UTM projection. I went through and right clicked a few of the feature classes and changed the projection to the State Plane in the same manner that I did the second. This time they shot off the screen as if I had defined the projection against what the file was created in. This is making me wonder if changing the XML actually did have some effect on the spatial reference and it just needed to be manually set because there was still something in the XML that I missed. I'm just a little lost since the menu that's supposed to show what the feature class is actually projected in, lets you change what it says its projected in, seemingly defeating its purpose.
I'm just hesitant to move forward working with this data, since I'm not sure its been completely corrected to the core and I don't want to run into errors in the future because of it. Can anyone offer any insight? Thank you!