Antalya 26.3: Fix empty partition_key and sorting_key in system.tables for Iceberg tables without data snapshots

[il9ue]

Closes #1235.

Summary

SELECT partition_key, sorting_key FROM system.tables returned empty strings for Iceberg tables that had no data snapshot. This was reliably observable for tables accessed via the Glue catalog (since Glue's metadata_location more frequently points at a snapshot-free metadata file), but also reproduced for any empty Iceberg table regardless of catalog (REST, Glue, or direct IcebergS3).

Root cause

IcebergMetadata::partitionKey() and IcebergMetadata::sortingKey() (introduced in #959, refined in #1026, ported to 25.8 in #1095) gated their work on the existence of a data snapshot:

auto [actual_data_snapshot, actual_table_state_snapshot] = getRelevantState(context);
if (!actual_data_snapshot)
    return std::nullopt;

This is semantically wrong. Partition spec and sort order are table-level properties recorded at the top level of the Iceberg metadata file (default-spec-id, default-sort-order-id, partition-specs, sort-orders) and exist independently of whether any data snapshot has been written. Code inspection of getState() confirms that actual_table_state_snapshot is fully populated (schema_id, metadata_file_path, metadata_version) regardless of whether a snapshot exists; only actual_table_state_snapshot.snapshot_id is std::nullopt, and that field is never read by getPartitionKey() or getSortingKey().

The gate was therefore dead-gating valid data. The fix removes it.

Change list

• src/Storages/ObjectStorage/DataLakes/Iceberg/IcebergMetadata.cpp
  • partitionKey(): removed the if (!actual_data_snapshot) early return.
  • sortingKey(): removed the if (!actual_data_snapshot) early return.
• src/Storages/ObjectStorage/DataLakes/Iceberg/Utils.cpp
  • getSortingKeyDescriptionFromMetadata(): added a defensive has() guard for sort-orders and default-sort-order-id. This is a pre-existing null-deref that was previously unreachable in practice (always behind the snapshot gate); after removing the gate, empty Iceberg V1 tables without sort-orders would have hit it. The guard mirrors the shape already present in getSortingKeyDisplayStringFromMetadata.
• Test added (see below).

No header changes. No StorageSystemTables.cpp changes — the existing null/exception guards added for #1210 (Glue segfault) remain untouched.

Behavior preservation

• REST path output for already-working (non-empty) tables is byte-identical, including the PR Follow-up for #959: proper spaces, remove nulls order info #1026 formatting (proper spaces, no NULLS FIRST/NULLS LAST).
• Tables with no partitioning still return empty string (getPartitionKeyStringFromMetadata already guards on missing partition-specs).
• Non-Iceberg data lakes (Hudi, Paimon, Delta Lake) are untouched — the virtual defaults in IDataLakeMetadata are unchanged.
• The defensive guards added for Segfault when getting info from Glue catalog #1210 in StorageSystemTables.cpp remain in place.

Out of scope

Glue's metadata_location pointer can lag schema-evolution events, which could cause partition_key / sorting_key to reflect a stale spec. This is orthogonal to the snapshot gate and is not addressed by this PR.

Test plan

New regression test reproduces Root Cause A without needing any catalog mock: creates an Iceberg table with a non-trivial partition spec and sort order, asserts system.tables.partition_key and system.tables.sorting_key are non-empty before any data is inserted.

Existing test_system_tables_partition_sorting_keys in tests/integration/test_storage_iceberg_with_spark/test_system_iceberg_metadata.py continues to pass with byte-identical output.

Changelog category (leave one):

• Bug Fix (user-visible misbehavior in an official stable release)

Changelog entry (a user-readable short description of the changes that goes to CHANGELOG.md):

Fixed system.tables.partition_key and system.tables.sorting_key returning empty strings for Iceberg tables that have no data snapshot, including all empty tables and (more frequently) tables accessed via the Glue catalog. Also added a defensive guard against Iceberg V1 metadata files missing sort-orders.

Documentation entry for user-facing changes

Not required — this is a bug fix to existing system.tables columns; no new user-facing surface.

CI/CD Options

Exclude tests:

• Fast test
• Integration Tests
• Stateless tests
• Stateful tests
• Performance tests
• All with ASAN
• All with TSAN
• All with MSAN
• All with UBSAN
• All with Coverage
• All with Aarch64
• All Regression
• Disable CI Cache

Regression jobs to run:

• Fast suites (mostly <1h)
• Aggregate Functions (2h)
• Alter (1.5h)
• Benchmark (30m)
• ClickHouse Keeper (1h)
• Iceberg (2h)
• LDAP (1h)
• Parquet (1.5h)
• RBAC (1.5h)
• SSL Server (1h)
• S3 (2h)
• S3 Export (2h)
• Swarms (30m)
• Tiered Storage (2h)