Weekly update in TiDB

Last week, we landed 48 PRs in the TiDB repository.

Added

• Add the HTTP API to query the DDL history
• Add two metrics gauges to show CPU and memory usage
• Propose the design to restore the SQL text from the AST tree
• Add the implementation phase framework of the cascades planner
• Add the ddl_reorg_batch_size variable to control the batch size of the DDL worker
• Add the gc_enable variable to enable or disable GC
• Add the transform phase framework of the cascades planner

Improved

• Speed up execution of unit tests for the statistics package
• Disable the global variable cache when running unit tests
• Do the DDL check for Create Index before putting it to the job queue
• Speed up execution of unit tests for the tikv package
• Improve compatibility when casting string as datetime
• Support Show Create Table for hash partitioned tables
• Support batchInsert of values for the Insert statement
• Improve the current greedy join reorder algorithm to leverage cost estimation
• Support Select/Insert for hash partitioned tables
• Support Truncate for hash partitioned tables
• Support Alter Table Add Partition for hash partitioned tables
• Support ? in Order By/Group By/Limit Offset clauses

Fixed

• Set correct concurrency for Projection below Aggregation
• Handle corrupted length parameters for the uncompress built-in function
• Fix failure of Grant in the ANSI_QUOTES SQL mode
• Fix incorrect date arithmetics with the negative interval parameter
• Fix the bug of cancelling Drop Index

Weekly update in TiSpark

Last week, we landed 4 PRs in the TiSpark repository.

Fixed

• Fix the unresolved column name when using table.column
• Fix the LockResolverTest error when sending requests using callWithRetry

Weekly update in TiKV and PD

Last week, we landed 22 PRs in the TiKV and PD repositories.

Added

• Add StoreNotMatch details
• Add the strcmp built-in function
• Add the week_day and the week_of_year built-in functions
• Add the compress, uncompress and uncompressed_length built-in functions
• Add the end_key support for raw_scan
• Add the concat_ws built-in function

Improved

• Replace FxHashMap with Hashbrown
• Make DeleteRange safe again

Fixed

• Reject transferring the leader to the recently added peers
• Change the merge match peers strategy

New contributors (Thanks!)

tidb:

• lonnng
• ziyi-yan

tikv:

• kg88
• AbnerZheng

docs-cn:

• beckxie