Introduction

Overview

Architecture

Eclipse plugin

Download

Report bug

Contact

About

The DwWorkflow is a server for running of refreshment process in datawarehouses. The ordinary datawarehouse consists of many sources and many targets and each target has at least one process which transforms data from different sources into this target. Each process can run after its sources have processed data. It means processes depend on their sources and DwWorkflow allows define these dependencies between processes and sources and then according to these dependencies can run these processes in parallel. The DwWorkflow also provides generating of mappings. The server can be accessed remotely using interface written in the interpreted language python. It is also possible to use a graphical application written as plugin in the eclipse platform (http://www.eclipse.org).

Main features

Flexibility

The usuall approach to create a workflow is to order procedures into some plan. But this approach is not flexible enough to adapt for different processing times and for parallel processing. The DwWorkflow offers this flexibility but it can be more complex compared to the forementioned approach. But this complexity is acceptable as this concept is approved on real projects.

Speed and reliability

The DwWorkflow is fast and reacts very quickly to the changes of the object states and to the user commands.

Server interface flexibility

The server is controled using console with python interface. So it is possible to use python to write complex and flexible scripts.

Generating of mappings

The DwWorkflow offers template engine for generating of mappings. There are some other tools which offer graphical way how to write mappings. But these tools look nice when they are used for simple mappings but when they are used for complicated mappings then we can get unreadable, unmaintainable and slow code. The approved way is to generate mappings. It allows the developer to write the core of the mapping and the tedious and repetitive parts of the mappings are generated. This approach when correctly used produces very fast and maintainable mappings.