Swarm: s3() + object_storage_cluster loses hostName() in remote aggregation (NOT_FOUND_COLUMN_IN_BLOCK)

[DimensionWieldr]

Describe the bug

Selecting hostName() with aggregation over s3(...) using SETTINGS object_storage_cluster='swarm' fails with:

Code: 10. DB::Exception: Not found column hostName() in block. There are only columns: count(): While executing Remote. (NOT_FOUND_COLUMN_IN_BLOCK)

Remote aggregation returns only count() and drops hostName(), so the initiator cannot complete GROUP BY host.

How to reproduce

With a configured swarm remote cluster and Parquet data under the warehouse (e.g. after creating/seeding an Iceberg table via PyIceberg):

SELECT hostName() AS host, count()
FROM s3(
    'http://minio:9000/warehouse/data/data/**/**.parquet',
    '<access_key>',
    '<secret_key>'
)
GROUP BY host
SETTINGS use_hive_partitioning = 1, object_storage_cluster = 'swarm';

Observed error:

Code: 10. DB::Exception: Not found column hostName() in block. There are only columns: count(): While executing Remote. (NOT_FOUND_COLUMN_IN_BLOCK)

Observed on 26.3.13.20001.altinityantalya.

Expected behavior

Each swarm worker should evaluate hostName() locally and return it with partial aggregates, so the initiator can group by host.