Berkeley DB Family

I am taking a 800MB CSV file and creating a berkeley db with about 50% of the information in the file. To do this I have created a custom class that contains a property for the columns of the CSV I need and a custom tuple binding (some of the columns aren't just simple types, although even with the custom tuple a single string "id" is stored which references a static external enumeration). Each row is instantiated as the class and put into the db as the value with one of the fields being used separately as the key.

The on-disk size of the Berkeley DB when all rows are imported is around 3GB (ultimately about 6x the size of the source data being stored)! Is this normal, or am I configuring something which is increasing my memory usage?

Thanks!