What pingcap/tidb shipped
Written by FoxPlug from the repository's own releases; not affiliated with Tidb. Paste your own repository in the box below and FoxPlug writes the same weekly update for your product. This page is an automatic summary of the releases, pull requests and commits in github.com/pingcap/tidb; Tidb did not write it and does not use or endorse FoxPlug. Every line links to the change it describes.
Get a weekly update like this for your product, free
Week of October 5, 2026
What shipped
- The parser now accepts CREATE VIEW... WITH CHECK OPTION without CASCADED or LOCAL, treating it as CASCADED like MySQL does. Pull request #71248
- SELECT INTO OUTFILE now honors LINES STARTING BY, so the prefix appears in the exported file. Pull request #71370
- Fixed JSON double formatting inside arrays and objects and for non-ASCII path keys. Pull request #71830
- CREATE MATERIALIZED VIEW and CREATE MATERIALIZED VIEW LOG now roll back the transaction if maintenance-info cleanup fails. Pull request #71816
- ANALYZE now derives NDV from schema-guaranteed uniqueness for columns and indexes kept distinct by a single-column unique key. Pull request #71721
- LOAD STATS now keeps FM sketches, so a later ANALYZE of one partition no longer merges a wrong global NDV. Pull request #71769
- Dumpling and Lightning now redact storage credentials such as access key, secret key and session token from logs and error messages. Pull request #71844
- IMPORT INTO now retries transient Aliyun OSS credential fetch failures from the instance metadata service. Pull request #71843
- ADD INDEX, ADD PRIMARY KEY and MODIFY COLUMN now require global sort when security enhanced mode is enabled on the nextgen kernel. Pull request #71836
- The coprocessor now skips empty TiDB nodes that only publish server info for schema sync, such as standalone BR. Pull request #71404
Why it matters
This week brought correctness fixes for views, exported files and materialized view cleanup, plus better statistics handling so NDV estimates survive LOAD STATS. It also tightened credential handling in storage logs and made IMPORT INTO and DDL more reliable under transient failures and security enhanced mode.
Changelog entry
- Parser: support CREATE VIEW... WITH CHECK OPTION without CASCADED or LOCAL Pull request #71248
- Executor: honor LINES STARTING BY in SELECT INTO OUTFILE Pull request #71370
- Types: fix JSON double formatting in containers and non-ASCII path keys Pull request #71830
- Ddl: rollback transaction on MV cleanup failure Pull request #71816
- Statistics: derive NDV from schema-guaranteed uniqueness Pull request #71721
- Statistics: keep FM sketches across stats dump and load Pull request #71769
- Statistics: refactor FM sketch merging in ANALYZE Pull request #71784
- Objstore, dumpling, lightning: redact storage URLs in logs and errors Pull request #71844
- Objstore: retry transient OSS credential fetch failures in IMPORT INTO Pull request #71843
- Ddl: require global sort for add index when SEM is enabled on nextgen Pull request #71836
- Coprocessor: skip empty tidb nodes Pull request #71404
- Server, session: add connection lifecycle and ALTER DATABASE logs Pull request #69866
This week: WITH CHECK OPTION parses without CASCADED or LOCAL, LINES STARTING BY works in SELECT INTO OUTFILE, credentials are redacted in logs, and LOAD STATS keeps FM sketches for correct global NDV.
This week in the project: CREATE VIEW... WITH CHECK OPTION now parses without CASCADED or LOCAL, matching MySQL, and SELECT INTO OUTFILE honors LINES STARTING BY. JSON double formatting was fixed inside containers and for non-ASCII path keys, and materialized view cleanup now rolls back the transaction on failure. Statistics improved too: ANALYZE derives NDV from schema-guaranteed uniqueness and LOAD STATS keeps FM sketches, so a later ANALYZE of one partition no longer merges a wrong global NDV. Storage logs and errors now redact credentials, IMPORT INTO retries transient OSS credential fetch failures, and on the nextgen kernel ADD INDEX requires global sort when security enhanced mode is enabled.
Week of September 28, 2026
What shipped
- Fixed a SELECT privilege bypass where users without column privileges could read constant-only table data. Pull request #71490
- Enabled plan caching for non-prepared UPDATE statements containing COALESCE expressions with matching column types. Pull request #71710
- Limited index selection preference based on range count to prevent indexes with many equality/IN predicates from dominating Skyline pruning. Pull request #70768
- Added support for WITH DEFAULT NDVRATE in ANALYZE statements to reset saved NDVRATE values. Pull request #71699
- Made the --pd flag optional for Dumpling when working with premium keyspace clusters. Pull request #71696
- Added diagnostic-mode guards for PD and TiKV RPC clients to control which operations are allowed. Pull request #71676
- Fixed resource unit accounting for ADD INDEX temporary index merge phase in NextGen. Pull request #71650
- Updated client-go dependency to fix a cancellation race blocking RunLoop.Append and TiKV BatchCommands receive loops. Pull request #71679
Why it matters
This week includes security fixes for privilege checks, improvements to query planning and caching that reduce unnecessary work, and fixes for distributed index operations and client communication races. These changes address both correctness issues and performance concerns across authentication, query optimization, and infrastructure layers.
Changelog entry
- Planner: fix SELECT privilege bypass for constant-only table reads Pull request #71490
- Planner: cache numeric COALESCE updates with matching column types Pull request #71710
- Planner: limit equality/IN preference by range count Pull request #70768
- Parser: support WITH DEFAULT NDVRATE in ANALYZE Pull request #71699
- Dumpling: make --pd optional for premium keyspace clusters Pull request #71696
- Client: guard PD and TiKV RPCs in diagnostic mode Pull request #71676
- Ddl: account RU for add-index temp-index merge Pull request #71650
- Store: update client-go to fix batch response cancellation race Pull request #71679
This week: fixed SELECT privilege bypass, enabled plan caching for COALESCE updates, improved index selection logic, added diagnostic mode guards for RPC clients, and fixed client-go cancellation race.
New fixes and improvements in TiDB this week: resolved a SELECT privilege bypass vulnerability, enabled plan caching for non-prepared UPDATE statements with COALESCE, refined index selection in query planning to prevent unnecessary seeks, added diagnostic-mode guards for RPC clients, and fixed a client-go cancellation race that could block BatchCommands operations. These changes enhance security, query performance, and cluster stability.
Week of September 21, 2026
What shipped
- Released v26.3.18. Release
- Fixed multi-statement latency attribution in query-duration metrics so slow statements in a batch are recorded with their actual type rather than the final statement's type. Pull request #71584
- Extended RU v2 accounting to partition reorganization and column modification DDL jobs that rewrite data. Pull request #71516
- Fixed materialized view refresh and log purge schedules to evaluate in UTC instead of persisted session timezones, resolving DST ambiguity issues. Pull request #71616
- Upgraded Go toolchain to 1.25.14. Pull request #71560
- Fixed panics and data races in statement summary subsystem reads on evicted table and cumulative metric paths. Pull request #70749
- Fixed distributed add-index to propagate table range scan errors instead of silently ignoring them. Pull request #71522
- Implemented materialized view log purge job execution and lifecycle management. Pull request #71118
- Added adaptive admission control for ordered LIMIT scans to avoid reading excess rows when a limit is present. Pull request #70252
- Fixed database read-only checks for resolved write targets in UPDATE, DELETE, and BATCH UPDATE statements. Pull request #71392
Why it matters
This week includes fixes for statement latency monitoring accuracy, distributed DDL error handling, and data race conditions that improve reliability. New features for materialized view scheduling and adaptive query execution provide better resource efficiency and data consistency guarantees.
Changelog entry
- Released v26.3.18 Release
- Added configuration gate for storage-class DDL and transition status queries to require explicit deployment opt-in Pull request #71551
- Fixed multi-statement latency attribution so each statement in a COM_QUERY batch is recorded with its correct type Pull request #71584
- Extended RU v2 accounting to partition and column modification reorganization DDL jobs Pull request #71516
- Changed materialized view refresh and log purge schedule evaluation to UTC timezone Pull request #71616
- Upgraded Go build toolchain to 1.25.14 Pull request #71560
- Fixed panics and data races in statement summary v1 and v2 evicted table reads and cumulative metric paths Pull request #70749
- Fixed distributed add-index to propagate table range scan errors Pull request #71522
- Implemented materialized view log purge job execution and lifecycle management Pull request #71118
- Added adaptive admission control for ordered LIMIT scans to prevent reading excess rows Pull request #70252
TiDB v26.3.18 and recent releases ship fixes for statement latency attribution, distributed add-index error propagation, and statement summary data races. Materialized view purge scheduling and adaptive LIMIT scans improve resource efficiency.
This week's TiDB updates focus on reliability and efficiency improvements. We fixed latency attribution in multi-statement queries to ensure accurate performance metrics, resolved data races in the statement summary system, and enhanced DDL resource accounting for data reorganization jobs. New features include materialized view log purge execution and adaptive admission control for LIMIT scans, which reduces unnecessary data reads and improves overall query efficiency.
Week of September 14, 2026
What shipped
- Lightning now accepts OSS URLs in data source configuration, enabling native Alibaba Cloud OSS access without requiring workarounds. Pull request #71316
- Client-go dependency updated to eliminate unnecessary warnings when decoding valid API v2 shared-lock wrappers. Pull request #71367
- User variable constant folding now applies to UPDATE, INSERT, and DELETE statements in addition to SELECT statements. Pull request #71386
- Global paging byte budgets now apply to existing sessions when changed via SET GLOBAL, enabling uniform tuning across connection pools. Pull request #71089
- Repeated preprocessing of prepared statements no longer incorrectly accumulates CTE consumer counts during schema changes. Pull request #70834
- Global sort merge operations now have bounded memory usage to prevent hotspot-reader blocks from consuming excessive resources. Pull request #70756
- NextGen now exposes a redacted global variables HTTP API for operators to inspect system variables without opening SQL connections. Pull request #71208
- Dumpling can now include values of STORED generated columns when dumping data for external system consumption. Pull request #71284
- RU v1 and v2 fields in slow logs and statement summaries are now unified into existing read/write RU fields. Pull request #71389
- Slow ADMIN CHECK now rejects partial multi-valued indexes that were previously bypassing validation. Pull request #70602
Why it matters
This week addresses multiple reliability issues across test stability, session handling, and resource management. Several backports fix authentication edge cases in the server, while planner improvements reduce incorrect query optimization decisions and new features expand OSS compatibility and monitoring capabilities.
Changelog entry
- Lightning: added support for OSS sources and preservation of OSS URIs in data source configuration Pull request #71316
- Client: updated client-go to fix unnecessary warnings when decoding API v2 shared-lock wrappers Pull request #71367
- Planner: extended read-only user variable constant folding to UPDATE, INSERT, and DELETE statements Pull request #71386
- Session: applied global paging byte budgets to existing sessions for uniform connection pool tuning Pull request #71089
- Planner: fixed CTE consumer count accumulation during repeated preprocessing of prepared statements Pull request #70834
- DDL: bounded global sort merge memory to prevent resource exhaustion from hotspot-reader blocks Pull request #70756
- Server: added redacted global variables HTTP API for NextGen operators Pull request #71208
- Dumpling: added support for dumping values of STORED generated columns Pull request #71284
- Executor: unified RU v1 and v2 fields in slow logs and statement summaries Pull request #71389
- Executor: reject partial multi-valued indexes in slow ADMIN CHECK operations Pull request #70602
This week: Lightning adds OSS support, prepared statements fix CTE accumulation bugs, and global paging budgets apply to existing sessions. Multiple test flakiness issues resolved.
This week's updates focus on reliability and feature completeness. Lightning now accepts OSS URLs natively, resolving data source configuration limitations. Session handling improvements fix issues where failed authentication could leave connections in invalid states. Query optimization fixes prevent CTE consumer counts from accumulating across prepared statement reprepares. Global paging byte budgets now apply uniformly to existing sessions. Multiple test flakiness issues stabilized.