The cloud-native orchestrator for the whole development lifecycle, with integrated lineage and observability.

dagster

Setting up my first sensor using `build_asset_reconciliation_sensor`. The sensor is queueing up some runs but they are failing and I believe it has something to do with partitioning and how the assets are defined.

The error I'm seeing for all of them is: `dagster._core.errors.DagsterInvariantViolationError: Cannot access partition_key for a non-partitioned run`

In the sensor cursor I'm seeing the following:
```[770136, ["asset_1", "asset_2"], {"asset_2": "{\"version\": 1, \"subset\": [\"2019-10-01\", \"2019-11-01\"]}", "asset_1": "{\"version\": 1, \"time_windows\": [[1679184000.0, 1679270400.0]], \"num_partitions\": 1}"}]```
The error looks to occur when I access `context.partition_key`.  Do the partitioned jobs need to be able to handle multiple partitions in their run?

Should I be writing all my partitioned assets to use `context.asset_partition_keys_for_output()` and if so, would I be able to dynamically yield multiple `Output`s?

Or should `build_asset_reconciliation_sensor` be launching individual runs for each partition?

&gt; Or should `build_asset_reconciliation_sensor` be launching individual runs for each partition?
It should be launching individual runs for each partition