Pulling Data from CSV vs. pulling data from database

Reading, searching and closing a file every time the page is loaded is highly time consuming. Database systems are highly optimized for reading and writing huge amount of records, querying data, providing schemes, relations between the data etc.

What will happen if 10K users hit the taxonomy page? How you will handle the updates of a taxonomy or the insertion of a new one in the file? How you can do efficiently queries regarding the value saved in the CSV file? What will happen if you delete the taxonomy or if the file has 1 million records?

The proper way is to save it in a database.

You can read more here and here