If you look online for airflow tutorials, most of them will give you a great introduction to what Airflow is. They will talk about the ETL as a concept, what DAGs are, build first DAG and show you how to execute it. What is being often skipped is how your DAG’s tasks should exchange data.
This article assumes you have at least a basic understanding of what Airflow is and how it works.
If you look online for airflow tutorials, most of them will give you a great introduction to what Airflow is. They will talk about the ETL as a concept, what DAGs are, build first DAG and show you how to execute it.
What is being often skipped is how these tasks should exchange data.
One thing to keep in mind while building tasks is to remember that tasks have to be idempotent. Simplifying for our case, it means that if the execution with similar input parameters succeeds, they will produce the same result no matter how many times you call them. This might not be always obvious when combined with task inputs/outputs.

#airflow #data-exchange #python