Is there a way to import into Datablaze using some key identifier to avoid duplicates?
This isn't directly possible using the user interface.
You could use the Data Blaze API to insert your data row by row and check for an existing matching row before you adding each row. Would that work for you?
Could you tell me more about your workflow with this data? There may be other alternatives if I understood your use case better.