On Querying Data and Metadata in Multiversion Data Warehouse - Data Warehousing Design and Advanced Engineering Applications - page 221

section we present the technical architecture of
the MVDW system, an application for executing
content queries and an application for executing
metadata queries.

SHOW EVOLUTION OF
DIMENSION dimension_name
INSTANCE OF STRUCTURE
(table_name.attribute_name(value)
->
table_name.attribute_name(value)
-> ...
table_name.attribute_name(value))
IN VERSION version_name
[VERSION FROM start_date
[TO end_date]]

Technical Architecture

The MVDW system is composed of the MVDW
storage and six software modules (cf. Figure 3),
namely: (1) MVDW manager , (2) MVDW man-
agement interface , (3) content query manager ,
(4) content query interface , (5) metadata query
manager , and (6) metadata query interface . The
modules are implemented in Java and they connect
to the MVDW storage by means of JDBC.
The MVDW storage manages and stores meta-
data describing the MVDW , DW schema versions,
and DW instance versions. The MVDW manager
is responsible for managing DW schema and DW
instance versions. In particular, it is responsible
for: (1) deriving new DW versions, (2) execut-
ing schema change operations, instance change
operations, and fact instance modifications, (3)
creating new schema objects from scratch, (4)
loading data into DW instance versions, and (5)
transactional management of the MVDW . The
MVDW is managed by means of the MVDW man-
agement interface . The content query manager is
responsible for parsing and executing multiversion
queries and for returning their results (possibly
integrated) to the content query interface . The
content query interface is responsible for entering
MVQs for execution and presenting their results
to a user. It visualizes the results of MVQs as
texts and charts. Similarly, the metadata query
manager is responsible for parsing and execut-
ing metadata queries and for returning the results
for visualization. Metadata queries are entered
for execution and their results are presented to a
user by means of the metadata query interface .
The parser of metadata queries is build based on
the SableCC library. The visualization software
is based on JGraph and JGraphpad pro .

For example, the below query retrieves the
evolution of the instance of dimension Product
existing in base version R_March . In this ver-
sion, dimension Product is composed of two
hierarchically connected levels, i.e., Products and
Categories, such that Products → Categories , as
specified in clause INSTANCE OF STRUCTURE.
The dimension instance is defined on attributes
Products.prod_name and Categories.cat_name .
The instance has the following structure poro-
therm brick → vat 7% . Since the VERSION FROM
clause was omitted, the query will address all
existing DW versions. Notice that the dimension,
the level tables and their attributes could change
their names in other DW versions. In this case
their equivalents in other DW versions will also
be considered by the query.

SHOW EVOLUTION OF
DIMENSION Product
INSTANCE OF STRUCTURE
(Products.prod_name
('porotherm brick') ->
Categories.cat_name('vat 7%'))
IN VERSION 'R_March'

PROTOTYPE IMPLEMENTATION
OF MVDWQL

The feasibility of the proposed MVDWQL was
verified in the prototype MVDW system. In this