Click house/apache arrow 23.0.1

[chethan-64]

Rationale for this change

Upgrade Apache Arrow from 23.0.0 to 23.0.1 and forward-port the ClickHouse-specific patches that were previously carried on top of Arrow 23.0.0.

What changes are included in this PR?

• Restore ClickHouse compatibility APIs

  • raw_orc_reader
  • count-from-record-batch

• Forward-port ClickHouse patches:

  • fix-deadlock
  • remove-abort-in-logging
  • try-fix-data-race
  • fix-uninit-value-msan
  • allow_map_key_optional
  • Fix clang-tidy

Are these changes tested?

Yes.

Verified by building ClickHouse successfully against the updated Arrow 23.0.1 branch with all forward-ported patches applied.

Are there any user-facing changes?

No direct user-facing changes.
This update is required to move ClickHouse to Apache Arrow 23.0.1 while preserving ClickHouse-specific functionality and fixes.

[CLAassistant]

All committers have signed the CLA.

[chethan-64]

Could some one create branch ClickHouse/release/23.0.1 , so i can set my PR target.
I don't have permission to create a branch in Clickhouse/arrow fork .

[chethan-64]

Forward-ported all ClickHouse-specific patches from apply-clickhouse-patches-to-release-23.0.0 onto Apache Arrow 23.0.1 and verified that ClickHouse builds successfully against the updated Arrow branch.

[rschu1ze]

Forward-ported all ClickHouse-specific patches from apply-clickhouse-patches-to-release-23.0.0 onto Apache Arrow 23.0.1 [...].

Not true. I'm looking

• at https://github.com/ClickHouse/arrow/pull/74/commits (this PR), and
• at the output of cd contrib/arrow; git log -- in the ClickHouse main repository

and I see that some patches are missing.

[chethan-64]

I found and forward-ported the following patches from apply-clickhouse-patches-to-release-23.0.0 onto ClickHouse/apache-arrow-23.0.1:

• Restore ClickHouse compatibility APIs (GetRawORCReader, RecordBatchCountRows)
• fix-deadlock
• remove-abort-in-logging
• try-fix-data-race
• fix-uninit-value-msan
• allow_map_key_optional
• Fix clang-tidy
• remove debug table validation checks (PR remove debug table validation checks #70)

Note that the functionality from raw_orc_reader and count-from-record-batch is preserved via the Restore ClickHouse compatibility APIs commit:

*GetRawORCReader()
*RecordBatchCountRows()

These were carried forward together in a single compatibility commit rather than as separate cherry-picked commits.

I also updated contrib/arrow to the latest commit and verified that ClickHouse builds successfully against the updated Arrow 23.0.1 branch.

Could you please point me to the specific patches that are still missing? I want to make sure I carry over the complete ClickHouse patch set.