I have an assignment where I have to normalise a dataset and design a schema in MySQL. Two of the attributes are unique identifiers for each tuple (that is, they are each primary keys in their own right). How do I satisfy lossless decomposition? I'm thinking of making a table which just contains these two attributes, and to use only one of them as a foreign key for joins to all other tables.
There is a similar issue with some other attributes, specifically that there are attributes representing location as latitude and longitude as well as attributes that define the same location but with a different numbering system (it's a district-based co-ordinate system). Same question as above - how do I satisfy lossless decomposition for these attributes?
