Database Reference
In-Depth Information
a
b
c
ChildLevel
Name
ParentLevel
Name
LevelName
IdentAttributes
OtherAttributes
IdentAttributes
OtherAttributes
IdentAttributes
OtherAttributes
d
RoleName
1
LevelName
1
LevelName
2
Fact Name
RoleName
2
IdentAttributes
OtherAttributes
IdentAttributes
OtherAttributes
Measure
Measure: AggFct
e
g
h
f
Additive
Semiadditive +!
Nonadditive
/Derived
HierName
+
percentage รท
Fig. 4.1
Notation of the MultiDim model. (
a
) Level. (
b
) Hierarchy. (
c
) Cardinalities.
(
d
) Fact with measures and associated levels. (
e
) Types of measures. (
f
) Hierarchy
name. (
g
) Distributing attribute. (
h
) Exclusive relationships
a level has a set of
attributes
that describe the characteristics of their
members. In addition, a level has one or several
identifiers
that uniquely
identify the members of a level, each identifier being composed of one or
several attributes. For example, in Fig.
4.2
,
CategoryID
is an identifier of the
Category
level. Each attribute of a level has a type, that is, a domain for
its values. Typical value domains are integer, real, and string. We do not
include type information for attributes in the graphical representation of our
conceptual schemas.
A
fact
(Fig.
4.1
d) relates several levels. For example, the
Sales
fact in
Fig.
4.2
relates the
Employee
,
Customer
,
Supplier
,
Shipper
,
Order
,
Product
,and
Time
levels. As shown in Fig.
4.1
d, the same level can participate several
times in a fact, playing different
roles
. Each role is identified by a name and
is represented by a separate link between the corresponding level and the
fact. For example, in Fig.
4.2
,the
Time
level participates in the
Sales
fact
with the roles
OrderDate
,
DueDate
,and
ShippedDate
. Instances of a fact are
called
fact members
.The
cardinality
of the relationship between facts
and levels, as shown in Fig.
4.1
c, indicates the minimum and the maximum
number of fact members that can be related to level members. For example,
in Fig.
4.2
,the
Sales
fact is related to the
Product
level with a one-to-many
cardinality, which means that one sale is related to only one product and
that each product can have many sales. On the other hand, the
Sales
fact
is related to the
Order
level with a one-to-one cardinality, which means that
Search WWH ::
Custom Search