Data Normalization - Information Systems Reengineering, Integration and Normalization - page 346

the relation into three binary relationships may or may not exist. If a relation can be
decomposed into three relations, which implies that there is join dependency in the
relation as shown in the above example of relation SPJ. Otherwise, the relation can-
not be decomposed into three relations because there is no join dependency in the
relation.
The normalized 5NF is relation JS (J#, S#), relation PJ (P#, J#), and relation
SP (S#, p#). Therefore, for the three relations as a result of three decomposability
as shown in relation SPJ with JD, these three relations can be joined together into
original relation SPJ without information loss.
In summary, the UNF implies that the table tuples containing multiple values
and without key value. The 1NF implies that the table tuples in atomic value, that
is, single value, and with key value. The 2NF implies that all the relations are in
fully functional dependency and there is no partial functional dependency. The 3NF
implies that all the relations are in fully functional dependency, and there is no TFD.
BCNF implies that all the candidate keys in each relation are also determinant. They
are also called strong third normal form, that is, they are more restricted than 3NF.
The 4NF implies that all the relations do not have multiple multivalued dependen-
cies. In other words, the 2-decomposibility of multiple multivalued dependency in
a relation is decomposed into two single valued dependency relations. The 5NF
implies that all the relations do not have join dependency. In other words, the 3-de-
composibility of join dependency in a relation is decomposed into three relations
which can join together into anUNF relation with the join dependency.
In summary, there are steps in normalization as follows:

1. Decompose all data structures that are not 2D into 2D relations of segments
2. Eliminate any partial dependency
3. Eliminate any transitive dependency
4. Eliminate any remaining FD in which determinant is not a candidate key
5. Eliminate any MVD
6. Eliminate any JD that are implied by candidate keys (Fig. 9.10 )

In general, relations in a higher normal form implies the relations which are already
in a lower normal form. Data dependency are data constraints rules among data
fields. TFD is redundant information in the relations. Also, all normal forms can be
denormalized, except BCNF cannot be denormalized without losing information in
some cases.

9.2 Normalize XML Document Without Data Redundancy

Data redundancy in a database is abnormal because it will be difficult to be updated
and maintained, which needs to be normalized to eliminate the redundant data with-
out loss of information.
Potential data redundancy may occur when a data semantic in an XML docu-
ment denotes a many-to-many relationship. To remove potential data redundancy