3. Overview of Transformation Manager
Transformation Manager (TM) is a stand-alone Windows® or Linux® software suite of meta-data driven code generator programs (see Figure 1) that provides for the authoring, testing, and debugging of data transforms between virtually all types of data, whether XML, databases, flat files, Java™ classes or spreadsheets, with flexible deployment options in both J2EE™ and Microsoft®.NET architectures.
TM’s suite of tools provides a complete development environment to author, test and debug transforms rapidly. TM models and transforms are stored in a Meta-data Text Repository, which is fully compatible with version control systems and provides an ideal multi-user environment for the sharing and developing of all aspects of models and transforms.
Models can be acquired from all kinds of data. For example, XML models may be acquired from DTD or XSD definitions, or even an XML Instance. The meta-data models are fundamental to TM and are used in all areas of building transformations, for example, the relationships between data entities determine the meaning of the data structures and mapping operations, and the constraints that they hold specify the operations allowed. The metadata for relational databases and Java classes are also acquired by direct access from the data sources. Flat files, and virtually all other kinds of data, may be accessed using TM’s published Open Interface.
The design tool (see Figure 2) provides an interactive design environment for importing and exporting models. It allows the user to describe transforms between different types of data using a drag and drop paradigm and displays simple, humanly readable, high-level transformation syntax, “Simple Mapping Language” (SML). SML is a very simple, intuitive transformation syntax, and although for simple cases some consider a graphical representation best, ETL Solutions have adopted a transformation syntax, as it is very easy to learn, and most importantly, provides the most powerful and flexible solution to ETL tasks.
Extensive experience has shown that the process of developing solutions to complex data transformations or integration problems depends upon the transparency of the data structures involved and of the description of the mappings themselves. It is essential therefore, that the data models be presented completely and meaningfully to the developers and analysts working on a project, and that these can be interpreted with the help of humanly readable mapping statements.
