Currently when an ad-hoc Datum Export job is created, any source ID patterns given in the data configuration are resolved into actual source IDs. It might be better for the SQL query to resolve the appropriate source IDs at execution time, rather than re-writing the sources submitted with the job configuration with the resolved sources, as this can be surprising to users when viewing the job status (seeing the source IDs have changed from what they submitted).

If the sources are to be resolved at execution time, they must take into consideration the permissions of the actor (user or token) that submitted the job, to restrict the export to only the nodes/sources allowed by that actor.