Fix attempt to create table from table function by ianton-ru · Pull Request #1701 · Altinity/ClickHouse

ianton-ru · 2026-04-28T17:09:51Z

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 attempt to create table from table function

Documentation entry for user-facing changes

After rewrite query to object storage with using remote initiator ClickHouse tried to create table.
Call of create relies on fact of column definitions on start, but with remote initiator feature column definitions are added during query rewrite.

CI/CD Options

Exclude tests:

Regression jobs to run:

github-actions · 2026-04-28T17:10:44Z

Workflow [PR], commit [66a0ce1]

ianton-ru · 2026-04-29T08:04:50Z

@codex review

ianton-ru · 2026-04-29T08:09:25Z

PR #1701 audit (Altinity/ClickHouse)

AI audit note: This review was generated by AI (gpt-5.3-codex).

PR: Fix attempt to create table from table function
Scope: Skip configuration->create() in StorageObjectStorageCluster when the storage is constructed as a table function (!is_table_function added to the existing guard).

Confirmed defects

No confirmed defects in reviewed scope.

Coverage summary

Item	Detail
Scope reviewed	`StorageObjectStorageCluster` constructor and the `configuration->create` guard; all `std::make_shared<StorageObjectStorageCluster>(...)` call sites (`registerStorageObjectStorage.cpp`, `TableFunctionObjectStorage.cpp`, `TableFunctionObjectStorageCluster.cpp`, `DatabaseDataLake.cpp`); base vs `DataLakeConfiguration::create` / `DataLakeMetadata::createInitial`; PR integration changes in `tests/integration/test_storage_iceberg/test.py`, `tests/integration/test_s3_cluster/test.py`, and cluster config.
Categories failed	None
Categories passed	Ephemeral table-function path must not run datalake `createInitial` (fix adds `!is_table_function`); persistent engine path still calls `create` (`is_table_function == false` in `registerStorageObjectStorage.cpp`); `DatabaseDataLake` with `is_datalake_query == true` already skipped `create` before this PR—unchanged.
Assumptions/limits	Static review only (tests not executed in this environment); commit message typo (“attemppt”) is cosmetic and not a runtime defect.

Fault-category matrix (in-scope)

Category	Status	Outcome
Table function must not trigger `create` / `createInitial` after query rewrite supplies columns	Executed	Pass
Persistent `CREATE TABLE` / engine registration must still invoke `create`	Executed	Pass
`is_datalake_query` interaction (`DatabaseDataLake`)	Executed	Pass (pre-existing `!is_datalake_query` guard)
Concurrency / lock-order / lifetime around this change	Not applicable	Constructor-only change
Exception / partial-update rollback	Not applicable	No change to surrounding `try`/`catch`

Call graph (anchor)

TableFunctionObjectStorage*::executeImpl / TableFunctionObjectStorageCluster::executeImpl / registerStorageObjectStorage / DatabaseDataLake::… → StorageObjectStorageCluster::StorageObjectStorageCluster → guarded configuration->create → (for datalake configs) DataLakeConfiguration::create → DataLakeMetadata::createInitial.

Code reference (change under review)

    if (!is_table_function
        && !columns_in_table_or_function_definition.empty()
        && !is_datalake_query
        && mode_ == LoadingStrictnessLevel::CREATE)
    {
        configuration->create(
            object_storage,
            context_,
            columns_in_table_or_function_definition,
            partition_by,
            if_not_exists,
            catalog,
            table_id_
        );
    }

ianton-ru · 2026-04-29T08:12:27Z

@codex review

chatgpt-codex-connector · 2026-04-29T08:19:26Z

Codex Review: Didn't find any major issues. More of your lovely PRs please.

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

ianton-ru · 2026-04-29T08:28:22Z

Failed tests looks unrelated to current PR

zvonand · 2026-04-29T10:34:54Z

Please, do not write "Solved .." in changelog

It belongs to PR body indeed -- just put it someplace else next time

zvonand · 2026-04-30T08:52:57Z

Unit test fail is fixed in antalya-25.8 branch already.
03707_set_index_bad_get_null_bug is pre-existing
03212_variant_dynamic_cast_or_default seems to be flaky and not related
test_startup_without_zookeeper is also pre-existing
test_filesystem_cache -- looks like it hit a timeout

alsugiliazova · 2026-04-30T15:56:39Z

@zvonand stop merging prs before verification. I did not had a chance to look at it.

alsugiliazova · 2026-04-30T22:56:28Z

Verification report: Altinity/ClickHouse PR #1701

Conclusion

PR is merged. CI on head was red but every failure is a pre-existing flake or known broken job; no PR-caused regression. No follow-up required.

CI on head `66a0ce1c` — failures

PR test workflow

Check	Test FAIL	Class
`Integration tests (amd_asan, old analyzer, 4/6)`	`test_replication_without_zookeeper::test_startup_without_zookeeper` (×2)	Pre-existing flake — 104 fails / 9 PRs / 90d, since 2026-03-30
`Integration tests (amd_binary, 1/5)`	same	same
`Integration tests (arm_binary, distributed plan, 2/4)`	same	same
`Integration tests (amd_asan, old analyzer, 6/6)`	`test_storage_s3_queue::test_persistent_processing`	New occurrence (1/1); s3_queue suite, no diff overlap
`Integration tests (amd_asan, flaky check)`	`test_storage_iceberg::test_filesystem_cache[s3-1-5]`, `[s3-4-5]`, `[s3-5-5]`, `test_metadata_cache[s3-4-5]`	Pre-existing flake — 2 fails / 2 PRs / 90d each. Caught by dedicated flaky-check job.
`Stateless tests (amd_binary, ParallelReplicas, s3 storage, parallel)`	`03707_set_index_bad_get_null_bug`	Pre-existing flake — 17 / 10
`Stateless tests (amd_binary, old analyzer, s3 storage, DatabaseReplicated, parallel)`	`03212_variant_dynamic_cast_or_default`	Pre-existing flake — 11 / 8
`Stress test (amd_debug)`	`Server died` + cascade (`Killed by signal`, `Hung check`, `Fatal message`, `gdb signal`)	Pre-existing instability — 8 SIGABRT / 3 PRs / 30d, also on master (PR=0 commit `5c9d5236` 2026-04-29 and `35573c93` 2026-04-25)
`Unit tests (asan)`	`IOTestAwsS3Client.AppendExtraSSE{C,KMS}HeadersRead/Write` (4 unit tests)	Pre-existing broken unit tests — 35 fails / 6 PRs / 90d each on `antalya-25.8` base

Regression workflow

Check	Top failure	Baseline
`RegressionTestsRelease / Swarms / swarms`	swarms node-failure / network-failure scenarios	Known broken/flaky swarms suite

Regression DB on /PRs/1701/ builds (30d): 19 Fail / 11,524 OK ≈ 0.16%. Failures concentrated in swarms (10) and s3_export_part (5) — both already known flaky/broken, not caused by this PR.

Related to PR #1701's diff?

PR is a 25.8 bugfix for CREATE TABLE … AS <table_function> rejecting/handling the remote-initiator + cluster-fallback path correctly (sibling to #1608/#1638's design issue #3.1 in AUDIT_PR_1608.md).

Failing test	Diff overlap	Related?
`test_replication_without_zookeeper::test_startup_without_zookeeper`	none (replication startup)	No
`test_storage_iceberg::test_filesystem_cache / metadata_cache`	none (cache layer)	No
`test_storage_s3_queue::test_persistent_processing`	none	No
`03707_set_index_bad_get_null_bug`	none (set index NULL)	No
`03212_variant_dynamic_cast_or_default`	none (Variant cast)	No
`Stress test SIGABRT`	none — also crashes on master	No
`IOTestAwsS3Client.AppendExtraSSE*`	none (S3 SSE headers)	No
`Swarms` regression	none	No

No failure intersects the TableFunctionObjectStorageClusterFallback / validateUseToCreateTable code path touched by this PR.

alsugiliazova · 2026-04-30T22:56:59Z

Code Audit: PR #1701 — Fix attempt to create table from table function

AI audit note: This review was generated by AI (Sonnet 4.6).

PR: Fix attempt to create table from table function
Commit: 17f55b0be08 merged as 53a7a1f695c into antalya-25.8
Fixes: #1683 — TABLE_ALREADY_EXISTS on SELECT from Iceberg table function on 25.8

Confirmed defects

No confirmed defects in reviewed scope.

ianton-ru and others added 2 commits April 28, 2026 18:58

Fix attemppt to create from table function

17f55b0

Merge branch 'antalya-25.8' into bugfix/antalya-25.8/fix_wrong_create

66a0ce1

ianton-ru added antalya bugfix antalya-25.8 labels Apr 29, 2026

zvonand approved these changes Apr 29, 2026

View reviewed changes

zvonand merged commit 53a7a1f into antalya-25.8 Apr 30, 2026
370 of 396 checks passed

alsugiliazova mentioned this pull request Apr 30, 2026

Unexpected TABLE_ALREADY_EXISTS on SELECT from iceberg table function on 25.8 antalya #1683

Closed

alsugiliazova added the verified Approved for release label Apr 30, 2026

svb-alt added the antalya-25.8.22.20001 label May 1, 2026

CarlosFelipeOR mentioned this pull request May 4, 2026

Antalya 25.8: Backport upstream fixes for parallel replicas + RIGHT/JOIN chains LOGICAL_ERROR #1724

Merged

27 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Fix attempt to create table from table function#1701

Fix attempt to create table from table function#1701
zvonand merged 2 commits intoantalya-25.8from
bugfix/antalya-25.8/fix_wrong_create

ianton-ru commented Apr 28, 2026 •

edited by zvonand

Loading

Uh oh!

github-actions Bot commented Apr 28, 2026 •

edited

Loading

Uh oh!

ianton-ru commented Apr 29, 2026

Uh oh!

ianton-ru commented Apr 29, 2026

Uh oh!

ianton-ru commented Apr 29, 2026

Uh oh!

chatgpt-codex-connector Bot commented Apr 29, 2026

Uh oh!

ianton-ru commented Apr 29, 2026

Uh oh!

zvonand commented Apr 29, 2026

Uh oh!

zvonand commented Apr 30, 2026

Uh oh!

Uh oh!

alsugiliazova commented Apr 30, 2026

Uh oh!

alsugiliazova commented Apr 30, 2026

Uh oh!

alsugiliazova commented Apr 30, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Conversation

ianton-ru commented Apr 28, 2026 • edited by zvonand Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Changelog category (leave one):

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

Documentation entry for user-facing changes

CI/CD Options

Exclude tests:

Regression jobs to run:

Uh oh!

github-actions Bot commented Apr 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ianton-ru commented Apr 29, 2026

Uh oh!

ianton-ru commented Apr 29, 2026

PR #1701 audit (Altinity/ClickHouse)

Confirmed defects

Coverage summary

Fault-category matrix (in-scope)

Call graph (anchor)

Code reference (change under review)

Uh oh!

ianton-ru commented Apr 29, 2026

Uh oh!

chatgpt-codex-connector Bot commented Apr 29, 2026

Uh oh!

ianton-ru commented Apr 29, 2026

Uh oh!

zvonand commented Apr 29, 2026

Uh oh!

zvonand commented Apr 30, 2026

Uh oh!

Uh oh!

alsugiliazova commented Apr 30, 2026

Uh oh!

alsugiliazova commented Apr 30, 2026

Verification report: Altinity/ClickHouse PR #1701

Conclusion

CI on head 66a0ce1c — failures

PR test workflow

Regression workflow

Related to PR #1701's diff?

Uh oh!

alsugiliazova commented Apr 30, 2026

Code Audit: PR #1701 — Fix attempt to create table from table function

Confirmed defects

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

ianton-ru commented Apr 28, 2026 •

edited by zvonand

Loading

github-actions Bot commented Apr 28, 2026 •

edited

Loading

CI on head `66a0ce1c` — failures