We should add:

• A UDF that is able to read partition size information from an Iceberg table
• A UDF in a general-enough format that is able to delete data based on specific columns and given filtering.