The data I am using to build an address locator is structured in a standard relational database model structure. Some of the fields that I need to provide e.g. Left City or Place are stored as codes in my feature class that point to an external lookup table which stores the real text name.
I do not want to flatten the data as the geodatabase is used for other tasks, is updated fairly regularly and is a large dataset. Coded domains are not recognised by the address locator and simply return the original code in the search results, which is not very useful.
I thought about using views but unfortunately this is not possible using a file geodatabase - only when using SDE. Do I have any other options, without flattening the data using a python script? (This is what I am trying to avoid).