The MySQL team would like to thank everyone in the community who contributed bug reports, patches, pull requests, and continued feedback that became part of the MySQL 26.7 release.

As always, community contributions help make MySQL better for everyone. In this release, we received contributions across the MySQL Server Optimizer, Parser, Prepared Statements, Performance Schema, Data Types, DDL, upgrade tooling, utilities, packaging, and Connector/J.

Below are the contributions included in MySQL 26.7, grouped by category. Contributor names are shown in title case, and company affiliations are included where available.

Optimizer

  • #109601 – Improved partition pruning when INSERT rows explicitly set a non-NULL AUTO_INCREMENT value — Hope Lee
  • #110125 – Fixed incorrect results returned from indexed-column access — John Jove
  • #112198 – Corrected multiple-equality optimization that could degrade performance — Yicheng Wei
  • #112737 – Applied const-table conditions consistently when building range scans — Xingyu Yang (Tencent)
  • #113288 – Improved dynamic-range planning for LEFT JOIN queries — Hope Lee
  • #113304 – Corrected unexpected LEAST() results — Jinsheng Ba
  • #114571 – Improved handling of automatically created keys for materialized CTEs — Wang Zhengmao (Alibaba)
  • #116611 – Fixed user-defined histogram import failures for BIGINT data — Tianfeng Li
  • #116741 – Corrected incorrect query results involving InnoDB temporary tables — Jingqi Tian (Alibaba)
  • #117750 – Improved FORCE INDEX behavior for IS NOT NULL predicates — Kaiwang Chen
  • #117791 – Improved FORCE INDEX behavior for IS NULL predicates — Kaiwang Chen
  • #118057 – Optimized check_promoted_index processing — Alex Xing
  • #119198 – Fixed incorrect results with reverse index range scans — Somrak Monpengpinij
  • #119442 – Simplified ORDER BY optimization when a column is filtered by IS NULL — Kishore Kumar (Databricks via GitHub)
  • #119922 – Reduced excessive histogram background update warnings — Tony Chen (Amazon via GitHub)
  • #119995 – Corrected FIND_IN_SET() returning the wrong result — Jingqi Tian (Alibaba)
  • #120226 – Fixed an OverflowBitset capacity mismatch that could trigger a server crash — Yakir Gibraltar
  • #120616 – Improved chunking strategy for tables with composite primary keys — Kamil Holubicki (Percona)

Parser and SQL execution

  • #110962 – Fixed unexpected warnings when joining a table with a trigger — Shun Yi
  • #111559 – Fixed CTE view syntax handling — Vitali Vinahradski
  • #116084 – Corrected invalid join order hint syntax in expanded queries — Kaiwang Chen
  • #116222 – Aligned parser digest handling for statements with trailing semicolons — Kaiwang Chen
  • #119672 – Fixed event parsing during upgrades when the event was created in ANSI SQL mode — Vitali Vinahradski

Prepared Statements and Performance Schema

  • #115524 – Registered missing key cache mutex and condition instrumentation — Huaxiong Song (Alibaba)
  • #115943 – Removed trailing semicolon noise from STATEMENT_DIGEST() — Kaiwang Chen
  • #119236 – Fixed a SQL reprepare issue that could drive CPU to 100% — Alex Xing

Data Types, Collations, and Functions

  • #112112 – Fixed derived-table column resolution with ROLLUP — Hope Lee
  • #117543 – Returned the correct collation for temporal results — Dirkjan Bussink
  • #118053 – Corrected CAST(NULLIF(…) AS DATETIME) returning the wrong result — Ximin Liang
  • #118658 – Fixed NULL detection for DATE columns in derived tables after derived merge — Xingyu Yang (Tencent)
  • #119861 – Corrected row comparison with empty-set subquery results — Xingyu Yang (Tencent)

DDL, Upgrade, and Server Infrastructure

  • #113962 – Fixed offset checking for insert-position overflow — Dirkjan Bussink
  • #114187 – Corrected an ENUM-to-SET typo — Yongfei Ji
  • #114455 – Fixed build failures with WITH_INNODB_EXTRA_DEBUG enabled — Sam Bird
  • #115136 – Fixed AUTO_INCREMENT value handling when adding a column — Huaxiong Song (Alibaba)
  • #116878 – Reduced false sharing with a hardcoded cache-line size — Kaiwang Chen
  • #117983 – Reduced memory usage during data dictionary upgrades with many views — Huaxiong Song (Alibaba)
  • #118060 – Fixed insert behavior after a NULL-field import error — Feng Chuanheng
  • #118132 – Improved upgrade performance when altering the general log — Karry Zhang (Alibaba)
  • #118233 – Fixed DDL handling for invisible indexes — Alex Xing

Utilities, Shell, and Packaging

  • #112719 – Added a line break after each mysqldump database row — Michiel Beijen
  • #115021 – Suppressed expected warnings in mysql_tzinfo_to_sql — Daniël van Eeden
  • #118802 – Fixed postrm script directory-removal handling — Evgeniy Patlan
  • #119222 – Improved util.dumpInstance() chunking for partial primary keys — Kamil Holubicki & Sveta Smirnova (Percona)
  • #120043 – Improved MySQL Shell upgrade checks for cross-database foreign keys — Vidyadhar Chelluru

Connector/J

  • #120096 – Fixed CharacterEncoding handling in Connector/J connection URLs — Rahman Rejepov

Miscellaneous

  • #104169 – Refactored redundant code — Hiroyuki Sasagawa

Thank you to everyone who contributed to MySQL 26.7. Every patch, bug report, and piece of feedback helps make MySQL stronger, more reliable, and easier to use for the millions of developers, DBAs, and organizations that depend on it every day.