Antalya 26.3 Backport of #100404 - Fix bad cast exception when inserting into Iceberg table partitioned by year/month/day on Date column

[mkmkme]

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):

Fix logical error exception when inserting into an Iceberg table with a Date column partitioned by year, month, or day transforms (ClickHouse#100404 by @alexey-milovidov).

Documentation entry for user-facing changes

...

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)

[github-actions]

Workflow [PR], commit [243ee36]

[alsugiliazova]

Verification: PR #1770

Title: Antalya 26.3 Backport of ClickHouse#100404 — Fix bad cast exception when inserting into Iceberg table partitioned by year/month/day on Date column
Base: antalya-26.3 ← backports/antalya-26.3/100404
Head: 243ee366eafc94bb3973b46990003dc4023bf8b8
Size: +69 / -3 across 4 files · State: OPEN · Mergeable

Changes

Production fix (2 files):

• src/Storages/ObjectStorage/DataLakes/Iceberg/ChunkPartitioner.cpp — replaces the hard-coded DataTypeString result type with the real result_data_types[transform_ind] when executing the partition transform (root cause of the bad cast).
• src/Storages/ObjectStorage/DataLakes/Iceberg/Utils.cpp — extends getAvroType() to also map UInt8, Int8, Int16 to the correct Avro type.

New tests (2 files):

• tests/integration/test_storage_iceberg_with_spark/test_writes_create_partitioned_table.py → adds test_writes_date_column_with_time_transforms (2 storage × 3 transforms = 6 cases).
• tests/integration/test_storage_iceberg_with_spark/test_writes_with_partitioned_table.py → adds test_writes_spark_date_partition_by_time_transform (2 storage × 3 transforms = 6 cases). Regression test for issue Logical error: 'Bad cast from type DB::IColumn* to DB::ColumnVector<unsigned short>*' with Iceberg Date column ClickHouse/ClickHouse#86337.

PR-added tests — all GREEN

12 new parametrized cases × 3 integration jobs = 36 runs, 0 failures.

Job	New tests run	All status
Integration tests (amd_asan, db disk, old analyzer, 4/6)	12	OK
Integration tests (amd_binary, 5/5)	12	OK
Integration tests (arm_binary, distributed plan, 2/4)	12	OK

Coverage:

• toYearNumSinceEpoch(d), toMonthNumSinceEpoch(d), toRelativeDayNum(d) (ClickHouse-side write) — all pass on s3 + local.
• Spark-side PARTITIONED BY (year|month|day(c0)) then ClickHouse insert — all pass on s3 + local.

The new code path has clean positive coverage on all three architectures/builds.

CI overview (head commit)

• PR test workflow: 0 test_status='FAIL' rows in DB. All real test runs are OK or SKIPPED. GH-level rollup is fully green for the PR test workflow.
• Regression workflow: 27 success / 69 skipped / 4 failure (chronic baseline).
• One pending action_required job (queue/auth, not a real failure).

Regression-workflow failures (chronic baseline on antalya-26.3)

Suite	Fails	Verdict
Swarms (Aarch64 + Release)	227	chronic baseline (initiator out of disk space, etc.)
Parquet (Aarch64 + Release)	34	chronic baseline
S3Export partition (Aarch64 + Release)	20	chronic baseline
S3Export part (Aarch64 + Release)	14	chronic baseline

Identical fingerprint to sibling antalya-26.3 PRs (1767, 1762, 1759, 1748, 1747, …). No new failure modes introduced.

Caveat — partial frontport

PR lands on antalya-26.3 while companion features from antalya-26.1 are still being frontported in parallel. Some regression suites still rely on settings/CLI args not yet wired in this base; a final re-verify is recommended once the rest of the frontport bundle lands.

Verdict

Safe to merge.

• Newly added tests pass 100% (36/36 runs) across 3 integration jobs and all 6 parametrizations each.
• No PR-caused regression at the test level.
• All red checks are pre-existing chronic baseline failures shared with sibling antalya-26.3 PRs.

[alsugiliazova]

Audit: PR #1770 — Antalya 26.3 Backport of #100404 — Fix bad cast exception when inserting into Iceberg table partitioned by year/month/day on Date column

AI audit note: This review comment was generated by AI (Cursor agent, audit-review skill).

Confirmed defects

No confirmed defects in reviewed scope.