TOWARDS AN AUTOMATIC DATA MART DESIGN
Ahlem Nabli, Ahlem Soussi, Jamel Feki, Hanêne Ben Abdallah
Faculté des Sciences Economiques et de Gestion de Sfax, Sfax, Tunisie
Faïez Gargouri
Institut Supérieur d’Informatique et du Multimédia de Sfax, Sfax, Tunisie
Keywords: OLAP requirements, Data marts
generation, data warehouse generation, design methodology.
Abstract: The Data Warehouse design involves the definition of stru
ctures that enable an efficient access to
information. The designer builds a multidimensional structure taking into account the users requirements. In
fact, it is a highly complex engineering task that calls for a methodological support. This paper lays the
grounds for an automatic, stepwise approach for the generation of data warehouse and data mart schemes.
For this, it first proposes a standard format for OLAP requirement acquisition. Secondly, it defines an
algorithm that transforms automatically the OLAP requirements into data marts modelled either as star or
constellation schemes. Thirdly, it overviews our mapping rules between the data sources and the data marts
schemes.
1 INTRODUCTION
Current software tools for Data Warehouses focus
on meeting end-user proposals (e.g., Business
Objects, Impromptu, Oracle Warehouse Builder).
On the other hand, OLAP tools are dedicated to
multidimensional analysis and graphical
visualization of results. In addition, there are
products to assist the administrator in the
construction of Data Warehouse (DW) and Data
Mart (DM) schemes. However, with these tools, the
DW and DM schemes must be built beforehand and,
in most cases, manually. Consequently, this task can
be tedious, error-prone and time-consuming,
especially with the large volume and variation of
data sources.
In this paper, we propose an automatic approach
to
the design of DM schemes. As illustrated in
Figure 1, our design approach consists of four main
tasks: (1) acquisition of OLAP requirements
specified as two/n-dimensional fact sheets leading to
“semi-structured OLAP requirements”, (2)
generation of star /constellation schemes by merging
the semi-structured OLAP requirements, (3) DW
generation schema by fusion of DM schemes and (4)
mapping the DM to the data sources.
Within this approach for DW and DM schema
desi
gn, we propose a tabular format for OLAP
requirement acquisition. In addition, we define an
algorithm that transforms automatically the OLAP
requirements into DM modelled either as star or
constellation schemes. We outline mapping rules
between the data sources and the DMs schemes.
Finally, the generated DM schemes are merged to
construct the DW schema via a set of unification
rules.
The remainder of this paper is organized as
fol
lows. Section 2 defines the OLAP requirements
acquisition. Section 3 details the generation of DM
schemes. Section 4 presents the mapping between
the data sources and DM. Section 5 outlines some
unification rules. Section 6 overviews relevant
proposal for DW, summarizes our proposal and
outlines future work.
2 OLAP REQUIREMENT
ACQUISITION
The requirements in decisional analysis can be
formulated in various manners, and most generally
in natural language sentences that describe standard
requests. In our approach, where we aim at a
computer aided design, we propose to collect the
user requirements in a format familiar to the
decision makers, i.e., as structured sheets. As
illustrated in Figure 2, this generic structure defines
226
Nabli A., Soussi A., Feki J., Ben Abdallah H. and Gargouri F. (2005).
TOWARDS AN AUTOMATIC DATA MART DESIGN.
In Proceedings of the Seventh International Conference on Enterprise Information Systems, pages 226-231
DOI: 10.5220/0002520702260231
Copyright
c
SciTePress
OLAP
requirements
Acquisition
DM
generation
Semi-
structured
OLAP
requirements
DW
Ontology
Graphic acquisition of
OLAP requirements
DM schemes
+mapped elements
DW
generation
DW schema
Data sources
Data source Mapping
Figure 1: DW design starting from OLAP requirements
Data sources
OLAP
requirements
Acquisition
DM
generation
Semi-
structured
OLAP
requirements
DW
Ontology
Graphic acquisition of
OLAP requirements
DM schemes
+mapped elements
DW
generation
DW schema
Data sources
Data source Mapping
Figure 1: DW design starting from OLAP requirements
Data sources
Figure 1: DW design starting from OLAP requirements
the fact to be analyzed and its domain, its measures
and the analysis dimensions. We designate this
structure by the acronym 2D-F sheet for Two-
Dimensional Fact sheet.
With this format, the user OLAP requirements
can be viewed as a set of 2/nD-F sheets, each
defining a fact and two/n analysis dimensions. To
analyze a fact with n (n>2) dimensions, we may
need to use several 2D-F sheets simultaneously or
hide one dimension at a time to add a new one to the
sheet to obtain nD-F sheet.
This format is privileged because it is familiar
and intuitive to decision makers and the simplest
visualization output to produce by programmers
(Lehner, 1998) ( Bonifati, 2001) .
The OLAP acquisition module (see Figure1)
uses an ontology specific to the application domain
and specialized for the DW: it supplies the basic
elements of OLAP specification, better assists the
user in specifying his/her needs and helps to avoid
certain ambiguities (due to synonyms) in names of
facts, dimensions, measures, etc.
The output of the OLAP acquisition phase is a
set of sheets defining by a set of elements: facts to
be analyzed, measures, dimensions of fact analysis,
attributes of dimensions, etc. These specified
requirements, called semi-structured OLAP
requirements, are the input of the next module.
Example: Figure 3 shows a 2D-F sheet that
analyzes the SALE fact referring to the commercial
domain. The measure Qty is recorded according to
the dimensions Client and Date.
MonthYearDate
City
Region
Client (Name, First-name)
SALE
( Qty)
MonthYearDate
City
Region
Client (Name, First-name)
SALE
( Qty)
Date
Hierarchy
Date
Dimension
Fact measureFact measure
Client
Dimension
Client
Hierarchy
Commercial
Domain
name
weak
attributes
3 DM SCHEMA GENERATION
A DM is subject-oriented. It is characterized by its
multidimensional schema made up of facts measured
along analysis dimensions.
Our approach aims at constructing the DM
schema starting from OLAP requirements specified
as a set of 2/nD-F sheets. Each sheet can be seen as a
partial description (or a multidimensional view) of a
DM schema. Consequently, for a given domain the
complete multidimensional schemes of DMs are
derived from all the sheets specified with the
acquisition module. For this we have defined a set of
algebraic operators to derive automatically the MD
schema (Nabli, 2005).
« Selection Name »
…….…….
P2.Vi
P2.V2P1.V2
P2.V2
P2.V1P1.V1
…..D2.P12D2.P1 « Dimension
D1.PN
…..
P2.Vn…….P2.V
1
D1.P2
…..
P1.V2P1.V1D1.P1
« Dimension Name : D1»« FACT NAME : F »
«M
f
1
, M
f
2
, M
f
3
,…. M
f
m
»
name : D2 »
…..D2.P12D2.P1 « Dimension
D1.PN
…..
P2.Vn…….P2.V
1
D1.P2
…..
P1.V2P1.V1D1.P1
« Dimension Name : D1»« FACT NAME : F »
«M
f
1
, M
f
2
, M
f
3
,…. M
f
m
»
« Selection Name »
…….…….
P2.Vi
P2.V2P1.V2
P2.V2
P2.V1P1.V1
name : D2 »
Dimension Hierarchies
Measures
Domain name
Values of
measures
Figure 2: Generic structure of 2D-F sheet
D
i
D D
D
n
Hidden Dimensions
Figure 3: T1. 2D-F sheet for the SALE fact.
Fi
g
ure 3: T1. 2D-F sheet for the SALE fact
Figure 2: Gene
r
ic Structure of 2D-F Sheet
TOWARDS AN AUTOMATIC DATA MART DESIGN
227
This derivation is done in two complementary
phases according to whether we want to obtain star
or constellation schemes:
1- Generation of star schemes which groups sheets
referring to the same application domain and
describing the same fact. It then merges all the
sheets identified within a group to build a star.
2- Generation of constellation schemes which
integrates the star schemes (issued from the previous
phase) relevant to the same application domain and
which may have common dimensions. Two star
schemes with common dimensions could be
integrated to build a constellation.
In the remainder of this paper, we use the following
notation:
- Dim(s): the set of dimensions in an nD-F sheet s
- Hier(d): the hierarchy of a dimension d,
- Meas(s): the set of measures in an nD-F sheet s.
3.1 Star Schema Generation
To generate the star schemes, we use the following
algorithm:
Begin
1. Given t nD-F sheets analyzing f
facts belonging to m analysis domains
(m<=t).
2. Partition the t sheets into the m
domains, to obtain G
dom1
, G
dom2
,….., G
domm
sets of sheets.
3. For each G
domi
(i=1..m)
Begin
3.1. Partition the sheets in G
domi
by
facts into G
F1
domi
,…… , G
Fk
domi
(k<=f)
3.2. For each G
Fj
domi
(j=1..k)
Begin
3.2.1. For each sheet s G
Fj
domi
For each dimension d dim(s)
Begin
- Complete the hierarchy of d to
obtain a maximal hierarchy.
- Add an identifier Id
d
as an
attribute.
End
3.2.2. Collect measures
M
3.2.3. Create the structure of a fact
F for Fj with Mes
Fj
.
3.2.4. Collect dimensions
D
3.2.5. For each d Dim
Fj
Begin
- Determine hierarchies
UU
Fj
dhierhier = )(
Fj
domi
Gs
sd
d
)dim(
- Create the structure of a
dimension D for d with hier
Fj
d
.
- Associate D with F.
End
End
End.
In this algorithm, the t nD-F sheets are first
partioned into domains. This first step ensures that
each star schema is generated in one domain. In turn,
this will reduces the number of comparisons used in
the constellation schema generation phase (see
section 3.2). A star schema is constructed for each
fact (F
j
) in steps 3.2.2. to 3.2.5.
A hierarchy is called maximal if it cannot be
extended upwards or downwards by including
another attribute (Moody, 2000).
Example. Let us extend the previous SALE
example with two additional fact sheets in Figure 4.
The star schema resulting from applying our
algorithm is shown in Figure 5.
Note that Id
Date
, Id
Client
and Id
Prod
were added as
U
Fj
domi
Gs
Fj
smeases
= )(
im
= )dim(
U
fj
domi
Gs
Fj
s
Month
Quarter
YearDate
category
Product (unit price, prod-name)SALE
( Qty, Amount)
Month
Quarter
YearDate
category
Product (unit price, prod-name)SALE
( Qty, Amount)
Commercial
Semester MonthYearDate
Slice
Age
Client (Name, First-name)SALE
( Qty, Revenue)
Semester MonthYearDate
Slice
Age
Client (Name, First-name)SALE
( Qty, Revenue)
Commercial
Figure 4: T2 and T3 two sheets for the SALE fact.
Figure 4: T2 and T3 two sheets for the SALE fact
SALE
qty
revenu
amount
Id
Client
Client
Department
City
Age
Id
prod
Product
Sub-Category
SALE
Id
Client
Client
City
Region
Slice
Age
Product-Name
Id
Prod
Product
Category
Qty
Revenue
Amount
First-Name
Name
Unit-Price
Date
Day
Date
Quarter
Month
Id
Date
Semerter
Year
SALE
qty
Id
Client
Department
City
SHIPMENT
Id
Supplier
Supplier
City
Region
Qty
Amount
Social-Reason
SALE
qty
revenu
amount
Id
Client
Client
Department
City
Age
Id
prod
Product
Sub-Category
SALE
Id
Client
Client
City
Region
Slice
Age
Product-Name
Id
Prod
Product
Category
Qty
Revenue
Amount
First-Name
Name
Unit-Price
Date
Day
Date
Quarter
Month
Id
Date
Semerter
Year
Date
Day
Date
Quarter
Month
Id
Date
Semerter
Year
SALE
qty
Id
Client
Department
City
SHIPMENT
Id
Supplier
Supplier
City
Region
Qty
Amount
Social-Reason
SALE
qty
Id
Client
Department
City
SHIPMENT
Id
Supplier
Supplier
City
Region
Qty
Amount
Social-Reason
SALE
qty
revenu
amount
Id
Client
Client
Department
City
Age
Id
prod
Product
Category
Sub-Category
Date
Day
SALE
Id
Client
Client
City
Region
Slice
Age
Product-Name
Id
Prod
Product
Category
Date
Qty
Revenue
Amount
Quarter
Month
Id
Date
First-Name
Name
Unit-Price
Semerter
Year
Figure 5: Star schema built from the T1 , T2
and T3 sheets.
SALE
qty
revenu
amount
Id
Client
Client
Department
City
Age
Id
prod
Product
Category
Sub-Category
Date
Day
SALE
Id
Client
Client
City
Region
Slice
Age
Product-Name
Id
Prod
Product
Category
Date
Qty
Revenue
Amount
Quarter
Month
Id
Date
First-Name
Name
Unit-Price
Semerter
Year
SALE
qty
revenu
amount
Id
Client
Client
Department
City
Age
Id
prod
Product
Category
Sub-Category
Date
Day
SALE
Id
Client
Client
City
Region
Slice
Age
Product-Name
Id
Prod
Product
Category
Date
Qty
Revenue
Amount
Quarter
Month
Id
Date
First-Name
Name
Unit-Price
Semerter
Year
F
SALE
qty
Id
Client
Department
City
SHIPMENT
Id
Supplier
Supplier
City
Region
Qty
Amount
Social-Reason
Date
Day
Date
Quarter
Month
Id
Date
Semerter
Year
SALE
qty
Id
Client
Department
City
SHIPMENT
Id
Supplier
Supplier
City
Region
Qty
Amount
Social-Reason
Date
Day
Date
Quarter
Month
Id
Date
Semerter
Year
Date
Day
Date
Quarter
Month
Id
Date
Semerter
Year
igure 5: Star schema built from the T1 , T2
and T3 sheets.
Figure 6: S2 Star schema
.
Figure 7. Constellation schema built from the stars
in Figure 5 and 6.
Figure 7. Constellation schema built from the stars
in Figure 5 and 6.
Figure 5: Star schema built from the
T1, T2 and T3 sheets
Figure 6: S2 Star schema Figure 7: Constellation schema built
from the stars in Figure 5 and 6
ICEIS 2005 - DATABASES AND INFORMATION SYSTEMS INTEGRATION
228
attributes to identify the dimensions. In addition,
several attributes were added to complete the
dimension hierarchies. This addition was done in
step 3.2.1.1. of the algorithm.
3.2 Constellation Schema Generation
In the previous phase, we have generated a star
schema for each fact in the same analysis domain.
These latter have to be merged to obtain star/
constellation schemes. For this, we have defined a
similarity factor to measure the pertinence of
schemes to be integrated (Feki, 2004).
Given S
i
and S
k
two star schemes in the same
analysis domain, the similarity factor Sim(S
i
, S
k
) is
calculated on the basis of n and m which are the
number of dimensions in S
i
and S
k
respectively, and
p which is the number of their common dimensions.
In addition, to enhance the quality of the
integration result, we define a matrix of similarities
that measures the similarity between each pair of
multidimensional schemes. This matrix is used to
decide which schemes should be integrated first.
Since Sim(S
i
, S
k
) = Sim(S
k
, S
i
) this matrix is
symmetric.
Given n star schemes of the same analysis domain
S
1
, S
2
, …..,S
n
. Each schema is defined by a name; it
analyzes a fact and has a set of dimensions. The five
steps of DM schema construction are as follows:
a- calculate the matrix of similarities MS,
b- find all the occurrences of the maximum max
of MS,
c- construct a constellation by merging all
schemes with the maximum similarity max.
d- Re-dimension MS by :
- dropping rows and columns of the merged
schemes,
- adding one row and one column for the
newly constructed schema.
e- if <stopcondition> is satisfied then exit else
return to step a.
The stopcondition is a boolean expression, true if
either size of MS is equal to 1or all the values in MS
are lower than a threshold set by the designer.
Example. Let us extend the previous example
with the additional star S2, shown in Figure 6.
The similarity matrix for S1 and S2 contains the
single value Sim(S1,S2) = 0.75.
The constellation schema resulting from applying
the above five steps is shown in Figure 7.
4 DM-DS MAPPING
The DW is built from several data sources (DS)
while its schema is built from the DM schemes.
Thus, the DM schemes must be mapped to the DS
schemes. In our approach, the DM-DS mapping
adapts the heuristics proposed by (Golfarelli, 1998),
(Boehnlein, 1999), (Abell’, 2001) and (
Tryfona,
1999) to map each element of the DM schemes (i.e.,
fact, dimension …) to one or more elements of the
DS schemes (entity, relation, attribute, ….).
Our DM-DS mapping is done in three steps: first,
it identifies from the DS schema potential facts (PF),
and matches facts in the DM schemes with identified
PF. Secondly, for each mapped fact, it looks for DS
attributes that can be mapped to measures in the DM
schemes (PM). Finally, for each fact that has
potential measures, it searches DS attributes that can
be mapped to dimensions attributes in the DM
schemes (PD).
A DM element may be derived from several
identified potential elements. In addition, the same
element can be mapped to several identified
potential elements. It may also happen that a DM
element is different from all potential elements,
which might require OLAP requirement revision.
4.1 Fact Mapping
Fact mapping aims to find for each DM fact (Fd) the
corresponding DS elements. For this, we first
identify DS elements that could represent facts (PF).
Then, we confront the set of Fd with all identified
PF. The result of this step is a set of (Fd, PF) pairs
for which the measures and dimensions must be
confronted to accept or reject the mapping (see
section 4.4).
- Fact identification: Each entity of the DS
verifying one of the following two rules becomes a
potential fact:
+
<=
=
.)/(
);()(75.0
),(
otherwisepmnp
mnpnif
SSSim
ki
F1: An n-ary relationship in the DS with
numerical attribute with n>=2;
F2: An entity with at least one numerical
attribute not included in its identifier.
- Fact matching: An ontology is used to find for
each fact in the DM schema all corresponding
potential facts. In this step, we may encounter one
problematic case: a DM fact has no corresponding
PF. Here the designer must intervene.
Note that, when a DM fact has several
corresponding PF, all mappings are retained until the
measures and dimensions are identified.
4.2 Measure Mapping
For each (Fd, PF) determined in the previous step,
we identify the potential measures of PF and
confront them with those of Fd.
TOWARDS AN AUTOMATIC DATA MART DESIGN
229
- Measure identification: since measures are
numerical attributes, they will be searched within
potential facts (PF) and “parallel” entities; they will
be qualified as potential measures (PM). The search
order is the following:
1. A non-key numerical attribute of PF.
2. A non-key numerical attribute of parallel
entities to PF.
3. A numerical attribute of the entities related to
PF
by a "one-to-one" link first, followed by those
related to PF
by a "one-to-many" link.
4. Repeat step 3 for each entity found in step 3.
Note that, two entities E1 and E2 are “parallel” if
the set of entities related to E1 by a one-to-one link
is included in the set of entities related to E2 by one-
to-one links.
- Measure matching: given the set of potential
measures of each PF, we use an ontology to find the
corresponding measures in Fd. A DM measure may
be derived from several identified PM. The
identified PM that are matched to fact Fd measures
are considered the measures of the PF.
In this step, we eliminate all (Fd, PF) for which
no correspondence between their measures is found.
4.3 Dimension Mapping
This step identifies potential dimension attributes
and confronts them with those of Fd, for each (Fd,
PF) retained in the measure mapping phase.
- Dimension attribute identification: Each attribute,
not belonging to any potential measures and
verifying the following two rules, becomes a
potential dimension (PD) attribute.
D1: An attribute in a potential fact PF;
D2: An attribute of an entity related to a PF
via a "one-to-one" or "one-to-many" link
(Abello, 2001). The entity relationships
take into account the transitivity.
Note that, the order in which the entities are
considered determines the hierarchy of the
dimension attributes. Thus, we consult the attributes
in the following order:
1. An attribute of PF, if any.
2. An attribute of the entities related to PF by a
"one-to-one" link initially, followed by the
attributes of the entities related to PF
by "one-to-
many" link.
3. Repeat step 2 for each entity found in step 2.
- Dimension matching: given the set of PD
attribute, we use an ontology to find the
corresponding attribute in Fd. If we can match the
identifier of a dimension d with a PD attributes, this
later is considered as a PD associated to PF.
In this step, we eliminate all (Fd, PF) for which no
correspondence between their dimensions is found.
4.4 Validation Mapping
The crucial step is to specify how the ideal
requirements can be mapped to the real system. The
validation may also give the opportunity to consider
new analysis aspects that did not emerge from user
requirements, but that the system may easily make
available. When a DM has one corresponding
potential fact, the mapping is retained. Whereas,
when a DM fact has several corresponding potential
facts {(Fd, PF)}, the measures of Fd are the union of
measures of all PF. This is argued by the fact that
the identification step associates each PM with only
one potential fact. Therefore, all sets of measures are
disjoint. Multiple correspondences of dimensions are
treated in the same way.
5 DATA WAREHOUSE
GENERATION
In our approach, we have distinguished two storage
spaces: the DM and the DW which are designed in
two different models. The DMs have
multidimensional models, to support OLAP analysis,
whereas the DW is structured as a conventional
database. We found the UML class diagram
appropriate to represent the DW schema.
The DM schema integration is accomplished
through the DW generation module (see Figure 1)
that operates in two complementary phases:
1- Transform each DM schema (i.e. stars and
constellations) into an UML class diagram.
2- Merge the UML class diagrams. This merger
produces the DW schema independent of any data
structure and content.
Recall that a dimension is made up of hierarchies
of attributes. The attributes are organized from the
finest to the highest granularity. Some attributes
belong to the dimension but not to hierarchies; these
attributes are called weak attributes, they serve to
label results.
The transformation of DM schemes to UML class
diagrams uses a set of rules among which we list the
following five rules.
Rule1: Transforming a dimension d into classes.
Build a class for every non-terminal attribute of
each hierarchy of d.
Rule2: Assigning attributes to classes
A class built from an attribute a gathers
- this attribute,
- the weak attributes associated to a, and
- the terminal attributes that are immediately
related to a and not having weak attributes.
ICEIS 2005 - DATABASES AND INFORMATION SYSTEMS INTEGRATION
230
Rule 3: Linking classes
Each class Ci built from attribute at level i of a
hierarchy h, is connected via a composition link
to the class C
i-1
, of the same hierarchy, if any.
Rule 4: Transforming facts into associations
A fact table is transformed into an association
linking the finest level classes derived from its
dimensions. Measures of the fact become
attributes of the association.
Note that all of the above four rules apply only to
non-date dimension. Rule 5 deals with date
dimension.
Rule 5: Transforming date dimension
A date dimension is integrated into each of its
related fact classes as a full-date, i.e., detailed
date.
6 RELATED WORK AND
CONCLUSION
There are several proposals to automate certain tasks
of DW design, c.f., (Cabibbo, 1998), (Golfarelli,
1998), (Hahn, 2000), (Peralta, 2003), (Sergio, 2003)
(Moody, 2000), (Marotta, 2002) and (Hahn, 2000).
Other works pertinent to automated DW design
mainly focus on the conceptual design, c.f.,
(Hüsemann, 2000) and (Phipps, 2002) which
generate the conceptual schema from an E/R schema
of the source database. However, these works do not
focus on a conceptual design methodology based on
users’ requirements and are, in addition, limited to
the sources described as E/R.
The work presented in this paper is a step towards
the automatic construction of DW schemes. More
precisely, it defined an approach that generates
automatically the DM schemes from precisely
specified OLAP requirements. Then, it showed how
the DW schema can be generated systematically.
We are currently optimizing the generation of DM
schemes from OLAP requirements. In addition, we
are verifying the completeness of the DM to DW
schema transformation rules. We are also working
on how to identify specific ontology for DM/DW
design.
REFERENCES
Abell’ A., Samios J., Saltor F., 2001. Understandding
Analysis Dimensions in a Multidimensional Object-
Oriented Model. DMDW 01, Interlaken, Switzerland,
June 4.
Boehnlein, M., Ulbrich-vom Ende, A., 1999. Deriving the
Initial Data Warehouse Structures from the
Conceptual Data Models of the Underlying
Operational Information System. DOLAP’99, USA.
Bonifati A., Cattaneo F., Ceri S., Fuggetta A., Paraboschi
S., 2001. Designing Data Marts for Data Warehouses.
ACM Transactions on Software Engineering
Methodology.
Cabibbo, LTorlone, R., 1998. A logical Approch to
Multidimensional Databases. EDBT’98, Spain.
Feki J., 2004. Vers une conception automatisée des
entrepôts de données : Modélisation des besoins
OLAP et génération de schémas multidimensionnels.
8
th
MCSEAI, 9-12 Mai, Souse-Tunisie.
Golfarelli M., Rizzi S., 1999. Designing the Data
Warehouse : Key Steps and Crucial Issues. Journal of
Computer Science and Information Manegement, vol.
2, n°3, p. 1-14.
Golfarelli, M., Maio, D., Rizzi, S.,1998. Conceptual
Design of Data Warehouses from E/R Schemes.
HICSS’98, IEEE, Hawaii.
Hahn K., Sapia C., Blaschka M., 2000. Automatically
Generating OLAP Schemes from Conceptual
Graphical Models. DOLAP’00, USA.
Hüsemann B., Lechtenbörger J., Vossen G., 2000.
Conceptual Data Warehouse Design. DMDW’00,
Sweden.
Kimball, R., 1996. The Datawarehouse Toolkit. John
Wiley & Son, Inc.
Lehner W., 1998. Modeling Large Scale OLAP
Scenarios. 6th International Conference on Extending
Database Technology (EDBT'98), Valence (Espagne),
23-27 Mars.
Marotta A., Ruggia R., 2002. Data Warehouse Design: A
schema-transformation approach. SCCC’2002. Chile.
Moody D., Kortnik M., 2000. From Enterprise Models to
Dimensionals Models: A Methodology for Data
Warehouse and Data Mart Design. DMDW’00,
Sweden.
Nabli A., Feki J., Gargouri F., 2005. Automatic
Construction of Multidimensional Schema from
OLAP Requirements. AICCSA’05, 3-6 January,
Cairo, Egypt.
Peralta V., Marotta A., Ruggia R., 2003. Towards the
Automation of Data Warehouse Design. Technical
Report. Universidad de la República, Uruguay.
Phipps C., Davis K., 2002. Automating data warehouse
conceptual schema design and evaluation. DMDW'02,
Canada,
Tryfona N., Busborg F., and Christiansen J. G. B., 1999.
StarER: A Conceptual Model for Data Warehouse
Design. Proceedings of the ACM DOLAP99
Workshop, Missouri, November 2-6.
TOWARDS AN AUTOMATIC DATA MART DESIGN
231