I have assets A and B and and asset job that selects both of them. I have some of asset A's partitions materialized and now run the asset job to materialize only new partitions. this apparently triggers materialization of the already materialized partitions of A again. Why is that?

When you run an asset job that selects multiple assets, Dagster will launch runs for each partition of the selected assets. If you have already materialized some partitions of asset A, and you run the asset job again, it may appear that Dagster is re-materializing the already materialized partitions of asset A. This could happen for a few reasons: 1. Dagster's Default Behavior: By default, when you materialize a partitioned asset, Dagster will launch a run for each partition you select. If you do not specify which partitions to materialize, Dagster may include all partitions, including those already materialized. 2. Asset Dependencies: If asset B depends on asset A and you are materializing asset B, Dagster might re-materialize asset A to ensure that asset B has the latest data. This behavior ensures that downstream assets are always working with the most up-to-date information from their dependencies. 3. Auto-Materialization Policies: If you have an auto-materialization policy set up for asset A, it might trigger re-materialization under certain conditions, such as when a parent asset updates or based on a schedule (cron job). 4. Partition Mappings: If there are partition mappings between assets A and B, and you are trying to materialize a partition of asset B that depends on a partition of asset A, Dagster might re-materialize the partition of asset A to satisfy the dependency for asset B. To avoid re-materializing already materialized partitions, you can explicitly select only the new partitions that you want to materialize when launching the asset job. This can be done through the UI or programmatically using Dagster's Python APIs. If you continue to experience unexpected re-materializations, it may be helpful to review the asset definitions, dependencies, and any auto-materialization policies to ensure they are configured as intended. Additionally, checking the partition mappings and ensuring that they are set up correctly can help prevent unnecessary re-materializations. If you need more specific guidance, please provide additional details about the configuration of your assets, their dependencies, and how you are triggering the asset job. This information will help in diagnosing the issue more accurately.