From Conventional to Multiversion Data Warehouse - Evolving Application Domains of Data Warehousing and Mining - page 61

Table 3. Data sharing scheme between versions of the data warehouse

School_ID School_Name Univ_Name Campus_Name S_location V1.0 V2.1

1 SICT KTH Kista 1 1 1

… … … … … … …

From Program, Universtyfact, Time
Where program joinwith university fact
And universityfact joinwith time
Group by programname, year;

We have also discussed the necessary metadata.
The metadata for the data warehouse lifecycle
is complex in nature, and it includes metadata
about operations, versions, schema sharing and
instance sharing. We also discussed querying the
multiversion DW. Conventional SQL cannot be
used to retrieve data from multiple data warehouse
versions, therefore extensions to SQL were also
described in this chapter.
Finally, we have presented a practical illustra-
tive example of application of the multiversion
data warehouse approach. The discussion and
illustration of the approaches that we provided
in this chapter are a useful aid to anyone deal-
ing with the versioning issues in practical data
warehouse lifecycle.

concluSIon

This chapter has focused on describing multi-
version data warehousing concepts as a solution
for handling shortcomings of conventional data
warehouses. Real-world changes in business
scenarios may subject to changes in operational
sources of the data warehouse. These changes
in operational sources may result in production
of inconsistent results from the data warehouse
(DW) due to the inability of the DW to adapt to
the new instances. Also, transformation of data
from changed operational sources to data ware-
house may result in information loss problems.
Similarly, conventional data warehouses don't
have the ability to handle schema changes.
The versioning data warehouse is an approach
to overcome deficiencies of conventional data
warehouses in handling change. According to
this approach, twelve possible operations are
possible which can result in changes to the data
warehouse. For each change which can result in
deficiencies, a new version of the data warehouse
(called child version) is created from the previous
version (called parent version) by a change adapta-
tion process. The relationship between parent and
child versions can be represented by a versioning
graph and the versions can share schemas as well
as instances.

referenceS

Bebel, B., Eder, J., Koncilia, C., Morzy, T., &
Wrembel, R. (2004). Creation and Management
of Versions in Multiversion Data Warehouse.
In Proceedings of ACM Symposium on Applied
Computing, (pp. 717-723), Cyprus.

Bebel, B., Krolinkowski, Z., & Wrembel, R.
(2006). Formal approach to modelling a multi-
version data warehouse. Bulletin of the Polish
Academy of Sciences , 54 (1), 51-62.

Blaschka, M., Sapia, C., & Höfling, G. (1999) On
Schema Evolution in Multidimensional Databases,
In Proceedings of the International Workshop
on Data Warehouse and Knowledge Discovery,
(LNCS 1676, pp. 153-164). Berlin: Springer.