tidb-server v1.0.2

• Optimize the cost estimation of index point query.
• Support the Alter Table Add Column ( ColumnDef ColumnPosition) syntax.
• Optimize the queries whose where conditions are contradictory.
• Optimize the Add Index operation to rectify the progress and reduce repetitive operations.
• Optimize the Index Look Join operator to accelerate the query speed for small data size.
• Fix the issue with prefix index judgement.

tidb-server v1.0.1

• Support canceling DDL Job.
• Optimize the IN expresssion.
• Correct the result type of the Show statement.
• Support log slow query into a seperate log file.
• Fix bugs.

tidb-server v1.0.0

• The SQL query optimizer:
  • Adjust the cost model
  • Analyze pushdown
  • Function signature pushdown
• Optimize the internal data format to reduce the intrium data size
• Enhance the MySQL compatibility
• Support the NO_SQL_CACHE syntax and limit the cache usage in the storage engine
• Refactor the Hash Aggregator operator to reduce the memory usage
• Support the Stream Aggragator operator

tidb-server 1.0 Pre-GA

• The SQL query optimizer:
  • Adjust the cost model
  • Use index scan to handle the where clause with the compare expression which has different types on each side
  • Support the Greedy algorithm based Join Reorder
• Many enhancements have been introduced to be more compatible with MySQL
• Support Natural Join
• Support the JSON type (Experimental), including the query, update and index of the JSON fields
• Prune the useless data to reduce the consumption of the executor memory
• Support configuring prioritization in the SQL statements and automatically set the prioritization for some of the statements according to the query type
• Completed the expression refactor and the speed is increased by about 30%

tidb-server rc4

• The SQL query optimizer refactoring:
  • Better support for TopN queries
  • Support the automatic choice of the of the Join physical operator based on the cost
  • Improved Projection Elimination
• The version check of schema is based on Table to avoid the impact of DDL on the ongoing transactions
• Support BatchIndexJoin
• Improve the Explain statement
• Improve the Index Scan performance
• Many enhancements have been introduced to be more compatible with MySQL
• Support the JSON type and operations
• Support the configuration of query prioritizing and isolation level
• Speed up DropIndex operation

tidb-server rc3

• The following features are added or improved in the SQL query optimizer:
  • Support incremental statistics
  • Support the Merge Sort Join operator
  • Support the Index Lookup Join operator
  • Support the Optimizer Hint Syntax
  • Optimize the memory consumption of the Scan, Join, Aggregation operators
  • Optimize the Cost Based Optimizer (CBO) framework
  • Refactor Expression
• Support more complete privilege management
• DDL acceleration
• Support using HTTP API to get the data distribution information of tables
• Support using system variables to control the query concurrency
• Add more MySQL built-in functions
• Support using system variables to automatically split a big transaction into smaller ones to commit

tidb-server rc2

• Query optimizer
  • Collect column/index statistics and use them in the query optimizer
  • Optimize the correlated subquery
  • Optimize the Cost Based Optimizer (CBO) framework
  • Eliminate aggregation using unique key information
  • Refactor expression evaluation framework
  • Convert Distinct to GroupBy
  • Support the topn operation push-down
• Support basic privilege management
• Add lots of MySQL built-in functions
• Improve the Alter Table statement and support the modification of table name, default value and comment
• Support the Create Table Like statement
• Support the Show Warnings statement
• Support the Rename Table statement
• Restrict the size of a single transaction to avoid the cluster blocking of large transactions
• Automatically split data in the process of Load Data
• Optimize the performance of the AddIndex and Delete statement
• Support “ANSI_QUOTES” sql_mode
• Improve the monitoring system
• Solve the problem of memory leak
• Fix Bugs

tidb-server rc2-preview

plan: only plans that have apply will add cache. (#2564)

tidb-server rc1

The TiDB RC1 is now released. See the following updates in this release:

TiKV:

• The write speed has been improved.
• The disk space usage is reduced.
• Hundreds of TBs of data can be supported.
• The stability is improved and TiKV can support a cluster with 200 nodes.
• Supports the Raw KV API and the Golang client.

Placement Driver (PD):

• The scheduling strategy framework is optimized and now the strategy is more flexible and reasonable.
• The support for label is added to support Cross Data Center scheduling.
• PD Controller is provided to operate the PD cluster more easily.

TiDB:

• The following features are added or improved in the SQL query optimizer:
  • Eager aggregation
  • More detailed EXPLAIN information
  • Parallelization of the UNION operator
  • Optimization of the subquery performance
  • Optimization of the conditional push-down
  • Optimization of the Cost Based Optimizer (CBO) framework
• The implementation of the time related data types are refactored to improve the compatibility with MySQL.
• More built-in functions in MySQL are supported.
• The speed of the add index statement is enhanced.
• The following statements are supported:
• Use the CHANGE COLUMN statement to change the name of a column.
• Use MODIFY COLUMN and CHANGE COLUMN of the ALTER TABLE statement for some of the column type transfer.

New Tools:

• Loader is added to be compatible with the mydumper data format in Percona and provides the following functions:
  • Multi-thread import
  • Retry if error occurs
  • Breakpoint resume
  • Targeted optimization for TiDB
• The tool for one-click deployment is added.

TiDB Beta4

beta4

plan: fix a panic bug. (#1914)