[Bug] dbt source freshness expected a timestamp but received a string about dbt-athena HOT 2 OPEN

Leaving this for anyone picking this up.
The first attempt, we could overwrite collect_freshness macro, e.g. here

Alternatively, but a bit more complicated, on the python side, we can overwrite this method in the adapter:

    def calculate_freshness_from_metadata(
        self,
        source: BaseRelation,
        macro_resolver: Optional[MacroResolverProtocol] = None,
    ) -> Tuple[Optional[AdapterResponse], FreshnessResponse]:
....

from dbt-athena.

From my side I can add than max(<ts_column>) as max_loaded_at doesn't work fine for dbt source freshness query. For large tables it comes to scan full data without partition predicate. It seems like for max, min functions in Iceberg tables don't using Iceberg metadata. From another side I can say that querying metadata scan almost nothing (few kb).

I propose override collect_freshness macro and depends on table_type and partition column use different queries for dbt source freshness command:
For non-partitioned Iceberg tables and all Hive tables use default query:

select
      max(<ts_column>) as max_loaded_at,
      {{ current_timestamp() }} as snapshotted_at
    from <source_table>

For partitioned Iceberg tables use the following query:

select
      max(data.<partitioned_by_ts_column>.max) as max_loaded_at
from "<source_schema>"."<source_table>$partitions"

from dbt-athena.