Database Reference
In-Depth Information
a
Client
ClientId
ClientName
ClientAddress
...
Account
Holders
Time
Account
Agency
Date
Event
WeekdayFlag
WeekendFlag
...
AccountNo
Type
Description
OpeningDate
...
AgencyName
Address
Area
NoEmployees
...
Balance
Amount
b
Client
ClientId
ClientName
ClientAddress
...
Holder
Time
Account
Agency
Date
Event
WeekdayFlag
WeekendFlag
...
AccountNo
Type
Description
OpeningDate
...
AgencyName
Address
Area
NoEmployees
...
Balance
Amount
Fig. 4.21
Two possible decompositions of the fact in Fig.
4.19
.(
a
)Creatingtwo
facts. (
b
) Including a nonstrict hierarchy
known). This could be represented by a measure in the
AccountHolders
fact in
Fig.
4.21
a and by a distributing attribute in the many-to-many relationship
in Fig.
4.21
b.
Another solution to this problem is shown in Fig.
4.22
. In this solution, an
additional level is created, which represents the groups of clients participating
in joint accounts. In the case of the example in Fig.
4.20
, two groups should
be created: one that includes clients
C1
,
C2
,and
C3
and another with clients
C1
and
C2
. Note, however, that the schema in Fig.
4.22
is not in the 1MNF,
since the measure
Balance
is not determined by all leaf levels, that is, it is only
determined by
Time
and
Account
. Therefore, the schema must be decomposed
leading to schemas similar to those in Fig.
4.21
, with the difference that in
Search WWH ::
Custom Search