Data Normalization - Information Systems Reengineering, Integration and Normalization - page 350

Normalized XML

˫ˠ˟

̋ ̌

̌

d R e

d_ref R e_ref d e

normalize

r1

r1

˸˲̅˸˹̇̂˸˾˸̌

m to n Cardinality

˷̅˸˹̇̂˷˾˸̌

ˡ̂̇˸ˍʳ ^d and ^e refer to elements d and e through ^dtodKey and ^etoeKey

Fig. 9.12 Normalized XML document in many-to-many cardinality

same information content under the root element as shown in Fig. 9.12 . Then,
we make all other XML elements with the same information content as XML
reference elements that point to E. As such, we preserve the m-to-n relationship
between them.

By the induction hypothesis, D0, D1, D2, …, D k all have the same information con-
tent. We have also proved that D k and D k + 1 have the same information content.
Thus, D0 and D k + 1 have the same information content and the proof is complete.
Example: Normalize XML document of Overlap Generalization.
Notice that b_ref and c_ref refer the specific attributes and/or sub-elements of
their own. The overlap parts (inherited from element “a” is normalized to be an
instance of element e) .
Example: Normalized XML document of m-to-n Cardinality