Authors:
Christian Lettner
and
Michael Zwick
Affiliation:
Software Competence Center Hagenberg GmbH, Austria
Keyword(s):
Data Analysis, Generic Data Models, Extract Transform Load, Data Warehouse.
Related
Ontology
Subjects/Areas/Topics:
Coupling and Integrating Heterogeneous Data Sources
;
Data Warehouses and OLAP
;
Databases and Information Systems Integration
;
Enterprise Information Systems
;
Legacy Systems
Abstract:
Industrial manufacturing companies produce a variety of different products, which, despite their differences in function and application area, share common requirements regarding quality assurance and data analysis. The goal of the approach presented in this paper is to automatically generate Extract-Transform-Load (ETL) packages for semi-generic operational database schema. This process is guided by a descriptor table, which allows for identifying and filtering the required attributes and their values. Based on this description model, an ETL process is generated which first loads the data into an entity-attribute-value (EAV) model, then gets transformed into a pivoted model for analysis. The resulting analysis model can be used with standard business intelligence tools. The descriptor table used in the implementation can be substituted with any other non-relational description language, as long as it has the same descriptive capabilities.