Designing Data Marts from XML and Relational Data Sources - Data Warehousing Design and Advanced Engineering Applications - page 60

Figure 1. Designing data mart schemas from XML and relational data sources

3. all methods consider that the candidate DM

built are of the same pertinence whereas
some of these DM may not be useful for
the decisional process; and

by XML documents in use; 2) it automates the
main design steps; 3) it assists the designer in
the choice of relevant multidimensional concepts
among those extracted by assigning to each one a
relevance level reflecting its analytical potential
for the decision-making; and 4) it keeps track of
the origin of each component in the generated DM
schema. This traceability is fundamental both to
automatically derive logical representations and
to define ETL processes.

4. the proposed approaches try to represent the

main DM properties at a conceptual level by
abstracting away details of an envisaged DW
implementation platform. “Unfortunately,
none of these approaches defines a set of
formal transformations in order to: (i) univo-
cally and automatically derive every possible
logical representation from the conceptual
model; or (ii) help the designer to obtain
the most suitable logical representation of
the developed conceptual model” (Mazón,
& Trujillo, 2008).

DATA MART DESIGN FOR
RELATIONAL AND xML SOURCES

Our design method is a bottom-up method for
DM that starts from a relational database source
and XML documents compliant to a given DTD.
Unlike existing approaches, ours is composed of
four steps (Figure 1) among which only the last
is manually conducted by the decision makers; in
this step, they adapt the automatically constructed
DM schemas for their particular needs.
As illustrated in Figure 1, our design method
starts with a data source pretreatment step to re-
solve the structural heterogeneity of the two types
of data sources. For the relational database, this

To overcome these problems, we propose a
bi-source method that builds DM schemas from
data sources either modeled as relational databases
or structured as XML documents compliant to a
given DTD. This method enjoys four main advan-
tages: 1) It overcomes the problem of absence/
obsoleteness of conventional documentation
( i.e. , E/A diagram, UML class diagram). In fact,
it exploits the recent version of the data source
extracted from the DBMS repository or described