Database Reference
In-Depth Information
Figure 5. The example screen shot of the metadata query interface showing the result of a version
query
An example screen shot that displays the results
of an object evolution query is shown in Figure
7. The query retrieves the evolution of hierarchy
H_Product
in dimension
Product
that originally
exists in base version from March. As we can
observe in Figure 7, the hierarchy in dimension
Product
is graphically displayed next to the DW
version that includes the hierarchy. For example,
in DW version
RV1
, hierarchy
H_Product
includes
only one level
Items
. In
RV2
, the hierarchy is
composed of two hierarchically connected levels,
namely
Products
→
Categories
.
Similarly as for a version query, a detail infor-
mation on the searched object in a selected DW
version is available after clicking on the node of
a version derivation tree.
The
MVDWQL
allows to: (1) query multiple DW
versions that differ with respect to their schemas,
(2) augment query results with metadata describing
changes made to the queried DW versions, and (3)
explicitly query metadata on the history of DW
changes and visualize their results. Two types of
queries on metadata are supported, namely: (1)
queries searching for DW versions that include an
indicated DW object and (2) queries retrieving the
history of the evolution of an indicated DW object.
The
MVDW
and
MVDWQL
have been successfully
implemented in a prototype system.
Since a few years data warehouses are applied
to the integration and analysis of complex data,
e.g., XML, spatio-temporal, and multimedia.
Moreover, semantic Web, knowledge bases,
ontology based systems, and various reasoning
technologies are becoming components of deci-
sion support systems. It is likely that these various
systems will also suffer from structural and content
changes in data sources and will face the same
problems as traditional data warehouses.
SUMMARY
In this chapter we presented the
MVDWQL
query
language, for the Multiversion Data Warehouse.
Search WWH ::
Custom Search