I have heard that it is possible to design the ultimate abstract
database that consists of one table for data and one table for
joins. The advantage to this design is when table structures
are constantly changing or new entities are constantly being
added. The "work" so to speak is transferred from changing
the table design to manipulating the data out of the database.
I have an assignment involving storing data for financial
products (derivatives, etc.). Since new products are contantly
being added with different characteristics, I am wondering
if this pure abstraction approach might be the way to go.
Can anyone explain to me what these two tables would look like,
their attributes and their join relationships?
