ninsbl opened a new issue, #58385:
URL: https://github.com/apache/airflow/issues/58385

   ### Description
   
   In data aware scheduling, it would be great if one could configure outlet 
`Datasets` / `assets` to set the `logical_date` of the consuming `DAGRun` to 
the `logical_date` of the triggering DAG and force clearing previous `DAGRuns`, 
similar to the possibilities in the `TriggerDagRunOperator`. This would be 
useful, since the calendar view uses the `logical_date` to present a 
time-overview of all `DAGRuns` and re-running upstream DAGs jus schedules new 
`DAGRuns`.
   
   ### Use case/motivation
   
   We are using Airflow to process time-series data that are made available on 
a ~ daily basis. Processing the data involves several DAGs depending on each 
other and we use both `TriggerDagRunOperator` and data aware scheduling 
(`Dataset` / `Assets`).
   
   With the `TriggerDagRunOperator` we can make sure that downstream DAGs have 
the same `logical_date` as the triggering DAG which provides us with useful 
insight in the calendar view. And we can clear possible previous runs, keeping 
the `DAGRun` time table clean and easy to search. When multiple other DAGs 
depend on an upstream `DagRun`, using the `TriggerDagRunOperator` becomes 
impractical and we switch to data-aware scheduling. 
   
   However, using `Dataset` / `Asset` for scheduling currently does not allow 
defining the `logical_date` and forcing to clear previous `DAGRuns`...
   
   I am willing to submit a PR, but I would need some guidance on where to 
start...
   
   ### Related issues
   
   _No response_
   
   ### Are you willing to submit a PR?
   
   - [x] Yes I am willing to submit a PR!
   
   ### Code of Conduct
   
   - [x] I agree to follow this project's [Code of 
Conduct](https://github.com/apache/airflow/blob/main/CODE_OF_CONDUCT.md)
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to