Database Reference
In-Depth Information
each relational symbol r i j ,j by the place symbol ( _ ) j (while replacing
¬
r i j ,j by
( _ ) j ) and by replacing r by unlabeled ( _ ) .
We will represent each operation q i by a composition of two operations
v i · q A,i , where q A,i O(r i 1 , 1 ,...,r i k ,k ,r q ) is the same expression as q i above
where r q is a relational symbol of the same type as r , and v i O(r q ,r ) is the
expression ( _ ) 1 ( y i )
¬
the place symbol
( _ )( y i ) where y i is the tuple of variables of the atoms
r ( y i ) and r q ( y i ) .
In the simplest case when χ i is a tautology r( x i )
r( x i ) , χ i is replaced in
S by q i =
q A,i =
v i =
1 r
O(r,r) (that is, by the identity mapping expression
( _ ) 1 ( x i )
( _ )( x i ) if x i is not empty; and by the expression ( _ ) 1
( _ ) for the
O(r ,r ) , otherwise).
4. ( Construct operad's operations )
Add the 'empty operation' operad 1 r
identity operation 1 r
O(r ,r ) in S , represented by the map-
ping expression ( _ ) 1 ( _ ) .
Return The set of abstract operad's operations denoted by an (mapping) arrow
M AB = S : A B
.
In what follows, we will use the term 'mapping-operad' from such a set of ab-
stract operad operations. The reason why we insert the identity 'empty operad's
operation' 1 r
S is that, as we will see, we can have the
Tarski's interpretations of database mappings where we do not transfer any infor-
mation from the source to the target database (i.e., with the empty information flux
of such a mapping). Consequently, 1 r is always an element of the set of operad's
operations obtained from a given SOtgd. As we will see in the algorithm of de-
composition of SOtgds, we will have cases where the mappings are generated with
SOtgd equal to the tautology r
O(r ,r ) in M AB =
r
.
Example 7
In the most trivial case when
M AB ={
r
r
}
(that is, when there is
no effective mapping between the schemas
) where r is a nullary predicate
symbol (i.e., the truth propositional letter in FOL, Definition 1 ), so that SOtgd is a
banal tautology r
A
and
B
M AB is always satisfied, we obtain that
r and, consequently,
MakeOperads(
M AB )
={
1 r }
. The particular cases for such trivial mappings are the
following:
M AA ={
r
r
}: A A
,
M A A ={
r
r }: A A
, and
Id A ={
r
r }: A A , the identity mapping ,
where
A = ( { r } , ) is the empty database schema. It is easy to verify that
MakeOperads {
r } ={
r
1 r }
.
Obviously, we cannot have the Tarski's interpretations for the empty schema
,
and we will need to define its interpretation in an ad hoc way in Sect. 2.4.1 dedicated
to R-algebras for the operads.
A
Search WWH ::




Custom Search