From Conventional to Multiversion Data Warehouse - Evolving Application Domains of Data Warehousing and Mining - page 49

warehouse, a versioning graph is proposed to be
used (Wrembel, 2005a).
A version derivation graph, as shown in Figure
4, can be integrated with a graphical interface
tool to provide visualization of versions and their
relationships. In the interface, each version is rep-
resented by a labeled node. The label of a node is
a name and a number that indicates a relationship
between versions e.g. in V 1.0, 1 , V is the name of
version, first 1 is the identification number of the
version, 0 is identification of its parent version
and last 1 shows the status of the version which
could be 0 or 1. Dotted lines indicate parent-child
relationships between versions i.e. tail of dotted
line is toward parent version and head is towards
child version. The version from which a version
is derived is called parent version and the derived
version is called child version.
In order to meet the initial decision support
requirements of an enterprise, a zero th version
of the data warehouse is developed. It is called
the root version and represented by root node in
versioning graph, with label V 0.0,1 . First 0 means
that it is the root version, and the second 0 means
it has no parent version and 1 represents active
status of version. Every version is valid for a
certain period of time and it is represented by 1 if
the version is valid or active, otherwise by 0 i.e.
inactive status or not available for use.

The versioning graph is embedded in a graphi-
cal interface, and it is available for administrators
to modify. Using this interface an administrator
can manage operations such as: delete a version
completely, insert a new version, change relation-
ship between versions and edit identification of
a version.

change Adaptation Process

Creation of a new data warehouse version depends
upon the change brought about by the evolution
operations. Here, we present a process called
change adaptation process, which can be used to
adapt changes in operational source of data ware-
houses. The sequence of steps is shown in Figure
5 and their brief explanation is as follows:

• Source change identification: the purpose
of this step is to identify changes in opera-
tional sources of the data warehouse. This
is done by comparing the schema before
change with the newly developed schema
or schema developed after the change. By
doing the comparison, changes in sources
can be identified.

• Version creation: in this step the admin-

istrator has to create a new version of the
data warehouse explicitly, in order to adapt

Figure 4. Version derivation graph (adapted from Bebel, 2004)