Database Reference
In-Depth Information
bookID
@title
rID
subID
@sub
id 1
'Algorithm Design'
id ε
id 13
CS
id 2
'Algebra'
id ε
id 22
Math
(a) Relation R book in I NL D OC ( T , D )
authID
bookID
nameID
afID
@nam
@af
id 11
id 1
id 111
id 112
Kleinberg
CU
id 12
id 1
id 121
id 122
Tardos
CU
id 21
id 2
id 211
id 212
Hungerford
SLU
(b) Relation R author in I NL D OC ( T , D )
Figure 15.3 Shredding of T into I NL S CHEMA ( D )
that lab ( n ) is starred in D , as well as the root of T .Then
μ
( n ) is the closest marked node
n that belongs to the path from the root to n .
In algorithm I NL D OC , and for the remainder of the chapter, we denote by id n the rela-
tional element representing the node n of a tree T .
Algorithm 15.2 I NL D OC ( T , D )
Require: D is a nested relational DTD, T conforms to D .
Ensure: I is a relational instance of the schema I NL S CHEMA ( D ).
for all marked nodes n of T do
let be the label of n ;
add to the relation R of I a tuple that contains elements
id n
ρ @ a ( n )
|
@ a
A D ( )
id μ( n )
|
if
= r
( n )= n , n is not marked.
id n
| μ
ρ @ a ( n )
( n )= n ,@ a
( n )) and
| μ
A D (
λ
n is not marked
where the identifiers and attributes values for each of the elements id n , id μ( n ) and
ρ @ a ( n ) coincide with the position of the attributes for id λ ( n ) , id μ( ) and A D (
( n ))
λ
of R .
end for
return I
It is then easy to verify that our Requirement 2 is satisfied: that is, given a DTD D and
Search WWH ::




Custom Search