Aggregation of tables registered in data collector during batch runs

[Xordak]

What's the problem this feature will solve?
Consideration of table aggregation during batch runs.

Describe the solution you'd like
Concatenate rows of tables over batch runs while adding an additional column corresponding to the run id in the collect_data method of batch_run.

Additional context
Hello,

In a project, I am using the tables attribute of the datacollector. I am logging completed "jobs" that only happen sometimes in a step and also does only concern a specific type of agent. I wanted to avoid adding unnecessary rows to the agent reporters that would log nonsense data for a "job" when there are no new jobs to log or when jobs do not have properties that the actual agents have. This keeps the agent reporter data compact.

However, looking at the batch_run source code, it seems that only model and agent reporter data is aggregated over runs. Since I was not able to use the batch_run function, I had to write a workaround with a model run method that saves tables to file. I was wondering if there is a deeper reason or a complication that is not that obvious?

Best Regards

[rht]

Sorry for the late response.

However, looking at the batch_run source code, it seems that only model and agent reporter data is aggregated over runs. Since I was not able to use the batch_run function, I had to write a workaround with a model run method that saves tables to file. I was wondering if there is a deeper reason or a complication that is not that obvious?

The table output just hasn't been worked out yet, and not due to a complication AFAICT. You are welcome to help integrate your custom workaround into batch_run, and we can discuss about the design.