Native Iceberg scans return NULL for omitted identity partition columns

[lyne7-sc]

Describe the bug

When an Iceberg table references data files that omit identity partition columns, Auron can return NULL instead of the partition values recorded in Iceberg metadata. This can occur after importing Spark/Hive-style partitioned files with add_files.

To Reproduce

1. Import Hive-style partitioned Parquet files into an Iceberg table using add_files. The files omit the partition column p.
2. Query select id, p from table_name with Auron enabled.

For a row in partition p='a', Auron returns (1, NULL) instead of (1, 'a').

Expected behavior

Return (1, 'a'), matching Iceberg's reader, which supplies the missing identity column from partition metadata.

Screenshots

Additional context