+----------------------------------------+-----------+--------------------------------------+
| Use case                               | Good fit? | Explanation                          |
+----------------------------------------+-----------+--------------------------------------+
| Every day, drop and recreate the users | Yes       | Assets are known before              |
| table and the user_recommender_model   |           | a run and are being updated          |
| model that depends on it               |           |                                      |
+----------------------------------------+-----------+--------------------------------------+
| Every day, read a file of user IDs     | No        | The set of assets to update          |
| and change the value of a particular   |           | is not known before running the job. |
| attribute for each user                |           |                                      |
+----------------------------------------+-----------+--------------------------------------+

Howcome “no” for the second use case? The assets are known ahead of time, i.e. a file structure that conforms to a known directory structure and file format https://docs.dagster.io/guides/dagster/enriching-with-software-defined-assets#when-should-i-use-software-defined-assets