12. Problems with normalisation

1. You need to be careful with trying to make data atomic. Just because you can split some types of data further, it isn't always necessarily the correct thing to do. For example, telephone number might contain the code followed by the number 01234 567890. It wouldn't be sensible to separate out these two items.

2. You can end up with more tables than an unnormalised database

3. The more tables and the more complex the database, the slower queries can be to run

4. It is necessary to assign more relationships to interact with larger numbers of tables

5. With more tables, setting up queries can become more complex