Question

I have this table and I am supposed to re-design it to remove all anomalies from it. I will not give the table but the dependencies.

The functional dependencies are
(Product, Store) -> Dept
Store -> Manager
Dept-> Assist
(Product, Store)->Price
Product-> Weight
Product-> Manufact
Manufact-> Manuloc

Product and Store are the keys in this relation

Ok so I have to remove the anomalies by breaking up the table and creating new ones and state what will be in each table. I am unsure how to do that.

Answer 1

Apply Armstrong’s axioms and the rules derived from them. (I’m sure this is in your textbook. Check the index.) For example, given these two functional dependencies (FDs):

(Product, Store) -> Dept
Dept -> Assist

you can apply the transitivity rule to determine an unstated FD:

(Product, Store) -> Assist

From the FDs you’re given, derive an irreducible set of FDs; this set determines your tables.

This example is from Date’s Introduction to Database Systems. (He gives a more complete treatment of it.) Given

A  -> BC
B  -> C
A  -> B
AB -> C
AC -> D

rewrite for right-hand singletons using Armstrong’s axioms:

A  -> B
A  -> C
B  -> C
A  -> B
AB -> C
AC -> D

The FD A -> B occurs twice; we can discard one without losing information. We can reduce AC -> D to A -> D, and we can eliminate AB -> C. And we can eliminate A -> C, leaving

A -> B
B -> C
A -> D

From this irreducible set of FDs, you’d derive the two tables

A -> BD
B -> C