Building a Simple ETL Workflow with Phoebe and Apache Airflow

Overview

Modern data engineering requires not only powerful tools for orchestration but also intuitive development environments. This is where Phoebe comes in—a modern Web IDE tailored for designing and managing Apache Airflow v3 workflows with ease.

In this blog post, we’ll explore how to use Phoebe to implement a simple ETL scenario, transforming employee data from a source table to a target table by applying a basic transformation: converting names to uppercase. Along the way, we’ll take advantage of Phoebe’s rich development perspectives, which streamline every aspect of the workflow design process.

Overview of Phoebe

Phoebe is a web-based IDE purpose-built for developing and managing workflows in Apache Airflow v3. It provides a seamless, integrated development experience with several powerful perspectives:

• Workbench: Central area for writing and editing your DAGs and other files.
• Embedded Airflow v3 Instance: A pre-integrated, fully functional Airflow v3 setup with direct access to the Airflow UI.
• Git Perspective: Easily manage your source code, commit changes, and work with repositories from within the IDE.
• Database Management: Interact with your databases visually—view tables, run queries, and manage schemas without leaving the browser.
• Integrated Terminal: Full access to a shell terminal for advanced CLI-based operations or debugging.

This IDE is designed for data engineers, developers, and workflow designers who want to rapidly build and deploy pipelines without constantly switching between tools.

Scenario Overview: Simple ETL with Airflow

To demonstrate Phoebe in action, we'll implement a classic Extract-Transform-Load (ETL) pipeline. The goal of this example is to:

• Extract data from a source database table named source_employees.
• Transform the data by converting the first_name and last_name fields to uppercase.
• Load the transformed data into a target table named target_employees.

This simple transformation highlights how Airflow can orchestrate database operations while giving us a chance to showcase the power of Phoebe's development environment.

Next Steps

In the next section of this blog, we’ll walk step-by-step through the implementation:

1. Setting up the project and Airflow DAG in Phoebe.
2. Creating source and target tables. Adding sample data to the source table.
3. Writing the transformation logic in Python using Airflow’s PythonOperator.
4. Testing and deploying the DAG via the embedded Airflow instance.
5. Verifying the result in the target table.

Follow the steps below or watch the recorded video.

Step-by-Step: Implementing the ETL in Phoebe

Start Phoebe instance locally

• To start a Phoebe instance locally we need a PostgreSQL instance. To simplify the setup, let's define a docker-compose.yaml file using the following content:

services:
  postgres:
    image: postgres:16
    container_name: postgres
    hostname: postgres
    environment:
      - POSTGRES_USER=postgres
      - POSTGRES_PASSWORD=postgres
      - POSTGRES_DB=postgres
    healthcheck:
      test: [ "CMD", "pg_isready", "-U", "postgres" ]
      interval: 10s
      retries: 5
      start_period: 5s
    restart: always
    networks:
      - phoebe_net

  codbex-phoebe:
    image: ghcr.io/codbex/codbex-phoebe:3.6.0
    container_name: phoebe
    environment:
      - PHOEBE_AIRFLOW_POSTGRES_USER=postgres
      - PHOEBE_AIRFLOW_POSTGRES_PASS=postgres
      - PHOEBE_AIRFLOW_POSTGRES_HOST=postgres
      - PHOEBE_AIRFLOW_POSTGRES_DB=postgres
    ports:
      - "80:80"
    depends_on:
      - postgres
    networks:
      - phoebe_net

networks:
  phoebe_net:
    driver: bridge

volumes:
  postgresql_data:
    driver: local

• Start the instances using the following command

docker compose up

• Once the created instance is started, access the Phoebe instance at http://localhost:80
• Login using the default credentials (user: admin, password admin)

Create a project using the starter template

• Open the Workbench perspective in Phoebe.
• At Welcome view, select Apache Airflow Starter template
• Set employees-etl for Project
• Set etl_employees_dag for File name
• Set etl_employees_dag for The id of the generated DAG
• Click on the Create button
• A starter project will be automatically generated for you

Register data source to the PostgreSQL instance

To allow the Phoebe instance to connect to the PostgreSQL instance, we need to define a data source file. This will allow us to inspect the PostgreSQL in the database perspective.

• Create a new folder with name db in the created project
• Create a file demo-postgre.datasource in folder db
• Add the content from this link

Define DB schema and add sample data

To define the structure of the tables and to fill the source table with a sample data, we need to create a several special files.

• Add new file called demo.schema (for the schema definition) to folder db with the content from this link
• Next, add a sample data file source_employees.csv with the content from this link
• Create a file demo-data.csvim with this content

Verify tables structure and data are created

• Save all files and click on Publish all button. After a few seconds these files will be registered and tables will be created. The source table should be filled with a sample data.
• Go to the Database perspective
• From the data sources drop down, choose the create data source demo-postgre
• Expand the public schema and inspect the tables
• Execute the following SQL statements to ensure that table source_employees has 1000 entries and that the target table is empty

  -- source_employees table
  SELECT * FROM source_employees;
  
  SELECT COUNT(*) FROM source_employees;
  
  -- target_employees table
  SELECT * FROM target_employees;
  
  SELECT COUNT(*) FROM target_employees;

  

Implement the ETL using DAG

• Go to the Workbench
• Open the DAG file etl_employees_dag.py and replace its content with this one
• Create folder tasks in dags folder
• Add the extract task logic
  • Create file extract_task.py
  • Add this content
• Add the transform logic
  • Create file transform_task.py
  • Add the content from this link
• Add the load task
  • Create file load_task.py
  • Add this content
• The structure should look like this
• Publish the project using Publish all button to apply these changes
• After a few seconds the changes will be applied

Trigger the DAG execution

• Go to the Airflow perspective
• Select Dags
• You should see the created DAG
• Click on the etl_employees_dag DAG
• Trigger a new execution using the Trigger button
• You should see a successful execution
• Verify ETL execution Now, let's verify that the target table contains entries with upper case for first name and last name
  • Go to Database perspective
  • Execute the following SQL statements

    SELECT COUNT(*) FROM target_employees;
    
    SELECT * FROM target_employees;

  • Table target_employees should have 1000 entries
  • All entries must be in uppercase

Congratulations, you have successfully implemented the ETL scenario!

GitHub repository for the project

The project we implemented can be found in this GitHub repository.
If you don't want to go through the blog steps, you can clone the repository using the Git perspective of Phoebe.
Once the project is cloned, you have to publish it.
Next you can explore and play with the content of the project.

Final Thoughts

This ETL pipeline, while simple, illustrates a foundational use case that many real-world workflows build upon.
Using Phoebe, we can:

• Rapidly prototype and test data workflows.
• Stay within a single, unified development environment.
• Easily manage code, data, and execution—all from the browser.

Whether you're just starting with Airflow or looking to streamline your pipeline development, Phoebe offers a productive and modern approach to orchestrating data workflows.