Enhancing ETL workflows

with TM1 DAG Factory

For businesses that routinely process large and complex datasets into TM1 models, leveraging Airflow as an ETL tool can significantly streamline daily data load workflows. Airflow's asynchronous engine facilitates parallel processing, accelerating data loading, while its web interface provides a visual means to manage workflows effectively.

Traditional Workflow Challenges

Traditionally, Airflow requires workflows to be defined in Python code. While this method is flexible, it introduces several complexities. Editing workflows necessitates the same rigorous source code control practices as software development, including version control and CI/CD pipelines, making the process slow and cumbersome. Moreover, the need for specific knowledge of Airflow can hinder standardization efforts and obscure the workflow process, affecting transparency.

The Advantages of a DAG Factory Approach

A dynamic approach to managing workflows, through the use of a DAG factory, presents many advantages. This method entails describing workflows as data in various formats, such as TM1 cubes, CSVs, JSON strings, XMLs, or SQL tables. Our open-source solution, the "TM1 DAG Factory," harnesses the power of TM1 cubes to define ETL workflows.

The Inception of TM1 DAG Factory

The TM1 DAG Factory was born out of a need to expedite development processes. Handling large and complex TM1 data models posed significant challenges, especially when modifications to a single module necessitated dealing with the entire data model. This approach was inefficient, increasing development times unnecessarily.

Our goal was to devise a solution capable of automatically generating partial datasets for specific development tasks, each containing only the data relevant to the task at hand. This system would also streamline data pruning, eliminating sensitive or unnecessary data (e.g., data older than one or two years). By storing workflow descriptions in TM1 cubes and utilizing tm1py to interface with the Airflow API, the TM1 DAG Factory enables the straightforward assembly of development datasets tailored to specific tasks.

Broad Utility and Advantages

The TM1 DAG Factory has proven invaluable not only in creating partial datasets for development but also in simplifying the management of daily data loads. It excellently supports parallel execution, enhancing efficiency in data loading. Workflow modularization is simplified, allowing for easy reuse and the assembly of new, even hierarchical or recursive workflows. Moreover, testing workflows is more straightforward.

Critically, once the descriptive structure of the data is understood, maintaining or developing ETL workflows does not require specific knowledge of Airflow or Python. This flexibility means the ETL solution can be swapped out, with the same data structures applicable across different ETL tools.


Adopting the DAG Factory approach for managing data loading workflows offers significant benefits, including reduced development time and enhanced transparency. The TM1 DAG Factory, stands as a testament to the power of innovative solutions in addressing complex data management challenges. It's an opportunity worth exploring for anyone looking to optimize their ETL processes.