Database Reference
In-Depth Information
sources. We focus on aggregates maintenance
with showing the limitations of the commercial
WMS (Warehouse Management System) and
propose some interesting solutions to overcome
them (Badri, Boufarès, Ducateau & Nefoussi,
2006; Badri, Heiwy & Boufarès, 2008). Our work
is tested exemplified on a real example which
have been generated using Oracle 11g Data Base
Management System (DBMS). Our results are
compared to Oracle ones. Finally, our future work
is given as a conclusion.
•
comp: C → P
(
C
)
is a function such that c
∉ comp(c)
for any c € C.
An element of C
is called a data source component. A com-
ponent c is called simple if
comp
(
c
) = ∅.
Any data source comes with a type system in
which data are interpreted. Each simple component
is interpreted in the system by a basic type like
integer, real, char, boolean. The type construc-
tors must allow us the interpretation of all the
components.
Let us now see how to represent relational,
object-relational and XML data using our formal
approach.
Relational and object-relational models
revisited:
The definition of a relational or object-
relational database can be done in two steps:
the first step allows us to describe the structure
of the objects of the real world to be stored and
the second step permits to define the objects to
represent.
In terms of the data models, the description of
the objects structures is called the database schema.
The data (the objects themselves), constitute the
database instance. Moreover, the specification of a
set of constraints, called integrity constraints give
the possibility to be restricted with the relevant
data of an application. An instance satisfying the
constraints is called a database state.
In the sequel, we mean by extended-relational
schema either a relational schema or an object-
relational schema. More details on the formal
representation of the relational and object database
are given in (Lellahi, 2002).
Let
REL
,
ATT
and
DOM
be three disjoined sets
describing intuitively the names of relations, at-
tributes and domains. Using
REL
,
ATT
and
DOM
,
an extended-relational DB schema is defined as
a triplet
S
= (
R
s
, ATT
s
,DOM
s
) associated to three
functions (
dom
s
,
att
s
, [[ ]]
s
):
dom
s
:
ATT
s
→ DOM
;
att
s
:
R
s
→ P
f
(
ATT
)
\ {
∅
}
; and [[ ]]
s
:
DOM
s
→ TY
PES
where
R
s
⊆
REL, ATT
s
⊆
ATT
and
DOM
s
⊆
DOM
and TYPES indicates various types defined
by the type system,
dom
s
associates to each at-
HETEROGENEOUS
DATA INTEGRATION: A
FORMAL APPROACH
In this section we present a theoretical approach that
allows the integration of various heterogeneous data
sources (DS) (Hamdoun, 2006; Hamdoun, Boufarès
& Badri, 2007; Badri, 2008). This approach takes
as an input an integration environment (a set of
DS and a set of relationships between them) and
returns a DW. It is generated by applying a set of
algorithms we have developed.
A Formal Approach for Modelling
Heterogeneous Data Sources
Let us recall at the first that if X is a set,
|X|
rep-
resents its cardinality and
P
(
X
) the whole of its
parts. The considered sets are finite sets, therefore
|X|
is a positive integer and
P
(
X
) is also finite.
A data source can be seen as a set of elements
accompanied by one or more relationships be-
tween them. However, we limit our work to the
consideration of only two relationships
. Thus, we
model a data source as follows:
A data source is a triplet (C, ref, comp)
where:
•
•
C is a set;
ref: P
(
C
)
→ P
(
C
)
is a binary relation such
that: X ref Y
⇒
|X|
=
|Y | ; and
Search WWH ::
Custom Search