How to write a schedule that launches a single run that fills a range of asset partitions?

[sryza]

Every night, I want to launch a single run that materializes a range of asset partitions.

[sryza]

You can accomplish this with a RunRequest that supplies the "dagster/asset_partition_range_start" and "dagster/asset_partition_range_end" tags:

@schedule(job=..., cron_schedule="@daily")
def my_schedule():
    ...
    return RunRequest(
        tags={
            "dagster/asset_partition_range_start": start_partition_key,
            "dagster/asset_partition_range_end": end_partition_key,
        }
    )

[geoHeil]

But this is not applicable for the DBT API? at least not by default?

[sryza]

@geoHeil that's correct: dbt assets are not partitioned by default, and after making them partitioned, you need to define how the partitions get translated to filters in the dbt statements.

[PatriciaAyuso]

@sryza @geoHeil
Here is a conversation about this #7683 (comment)

[ytoast]

Would this work for static partitions as well?

[sryza]

@ytoast it should

[garethbrickman]

To write a sensor or schedule that requests a run for every partition on every tick, see #15532

[AdrianTheopold]

I have an asset with a dynamic partition that is using context.partition_keys.
How can I write a schedule/sensor (or a general python script for that matter) that runs the asset creation for a large number of partitions, say 2000 in my case.

I am able to run backfills through the UI which is very inconvenient for a large selection of partitions. But other than that I don't see an option for creating such a backfill via a sensor or script.

$ dagster job backfill -j do_stuff_partitioned --partitions
does not work for such a large number of partitions.