I have a module with single applies to that is my list of transactions. This has various line items (essentially properties). I then have an import routine which updates these values, my question is can you have it so that it doesn't update if there's already something there so only does the blanks?
If the intersection of dimensions is represented in the import file, then the cell will be updated. In this context, you cannot load a transaction with, say, 5 properties/attributes configured in the import, and not update all 5.
You could create targeted imports where, even though the CSV/Text file contains all 5 attributes, the import profile you use only maps 1 or 2 of the 5 (so that you dont' disrupt the other 3).
Data import options might help also, as we can limit the changes to the contents of the import file. if you put line items (i.e. the different attributes) down a single column, the the resulting import CSV/TXT file could omit unchanged line items, allowing you to load a very targeted file. For example, for some transactions in the CSV might have all 5 attributes represented, while others only have 1-4... in this format, the missing intersections wouldn't be updated. (I think this will work, but would need to test it). The nuance here is that you have to know which specific items you want to update, which implies some kind of analysis or filter being applied the CSV file before loading it.