- Docs Home
- About TiDB
- Quick Start
- Develop
- Overview
- Quick Start
- Build a TiDB Cluster in TiDB Cloud (Developer Tier)
- CRUD SQL in TiDB
- Build a Simple CRUD App with TiDB
- Example Applications
- Connect to TiDB
- Design Database Schema
- Write Data
- Read Data
- Transaction
- Optimize
- Troubleshoot
- Reference
- Cloud Native Development Environment
- Third-party Support
- Deploy
- Software and Hardware Requirements
- Environment Configuration Checklist
- Plan Cluster Topology
- Install and Start
- Verify Cluster Status
- Test Cluster Performance
- Migrate
- Overview
- Migration Tools
- Migration Scenarios
- Migrate from Aurora
- Migrate MySQL of Small Datasets
- Migrate MySQL of Large Datasets
- Migrate and Merge MySQL Shards of Small Datasets
- Migrate and Merge MySQL Shards of Large Datasets
- Migrate from CSV Files
- Migrate from SQL Files
- Migrate from One TiDB Cluster to Another TiDB Cluster
- Migrate from TiDB to MySQL-compatible Databases
- Advanced Migration
- Integrate
- Maintain
- Monitor and Alert
- Troubleshoot
- TiDB Troubleshooting Map
- Identify Slow Queries
- Analyze Slow Queries
- SQL Diagnostics
- Identify Expensive Queries Using Top SQL
- Identify Expensive Queries Using Logs
- Statement Summary Tables
- Troubleshoot Hotspot Issues
- Troubleshoot Increased Read and Write Latency
- Save and Restore the On-Site Information of a Cluster
- Troubleshoot Cluster Setup
- Troubleshoot High Disk I/O Usage
- Troubleshoot Lock Conflicts
- Troubleshoot TiFlash
- Troubleshoot Write Conflicts in Optimistic Transactions
- Troubleshoot Inconsistency Between Data and Indexes
- Performance Tuning
- Tuning Guide
- Configuration Tuning
- System Tuning
- Software Tuning
- SQL Tuning
- Overview
- Understanding the Query Execution Plan
- SQL Optimization Process
- Overview
- Logic Optimization
- Physical Optimization
- Prepare Execution Plan Cache
- Control Execution Plans
- Tutorials
- TiDB Tools
- Overview
- Use Cases
- Download
- TiUP
- Documentation Map
- Overview
- Terminology and Concepts
- Manage TiUP Components
- FAQ
- Troubleshooting Guide
- Command Reference
- Overview
- TiUP Commands
- TiUP Cluster Commands
- Overview
- tiup cluster audit
- tiup cluster check
- tiup cluster clean
- tiup cluster deploy
- tiup cluster destroy
- tiup cluster disable
- tiup cluster display
- tiup cluster edit-config
- tiup cluster enable
- tiup cluster help
- tiup cluster import
- tiup cluster list
- tiup cluster patch
- tiup cluster prune
- tiup cluster reload
- tiup cluster rename
- tiup cluster replay
- tiup cluster restart
- tiup cluster scale-in
- tiup cluster scale-out
- tiup cluster start
- tiup cluster stop
- tiup cluster template
- tiup cluster upgrade
- TiUP DM Commands
- Overview
- tiup dm audit
- tiup dm deploy
- tiup dm destroy
- tiup dm disable
- tiup dm display
- tiup dm edit-config
- tiup dm enable
- tiup dm help
- tiup dm import
- tiup dm list
- tiup dm patch
- tiup dm prune
- tiup dm reload
- tiup dm replay
- tiup dm restart
- tiup dm scale-in
- tiup dm scale-out
- tiup dm start
- tiup dm stop
- tiup dm template
- tiup dm upgrade
- TiDB Cluster Topology Reference
- DM Cluster Topology Reference
- Mirror Reference Guide
- TiUP Components
- PingCAP Clinic Diagnostic Service
- TiDB Operator
- Dumpling
- TiDB Lightning
- TiDB Data Migration
- About TiDB Data Migration
- Architecture
- Quick Start
- Deploy a DM cluster
- Tutorials
- Advanced Tutorials
- Maintain
- Cluster Upgrade
- Tools
- Performance Tuning
- Manage Data Sources
- Manage Tasks
- Export and Import Data Sources and Task Configurations of Clusters
- Handle Alerts
- Daily Check
- Reference
- Architecture
- Command Line
- Configuration Files
- OpenAPI
- Compatibility Catalog
- Secure
- Monitoring and Alerts
- Error Codes
- Glossary
- Example
- Troubleshoot
- Release Notes
- Backup & Restore (BR)
- TiDB Binlog
- TiCDC
- Dumpling
- sync-diff-inspector
- TiSpark
- Reference
- Cluster Architecture
- Key Monitoring Metrics
- Secure
- Privileges
- SQL
- SQL Language Structure and Syntax
- SQL Statements
ADD COLUMN
ADD INDEX
ADMIN
ADMIN CANCEL DDL
ADMIN CHECKSUM TABLE
ADMIN CHECK [TABLE|INDEX]
ADMIN SHOW DDL [JOBS|QUERIES]
ADMIN SHOW TELEMETRY
ALTER DATABASE
ALTER INDEX
ALTER INSTANCE
ALTER PLACEMENT POLICY
ALTER TABLE
ALTER TABLE COMPACT
ALTER USER
ANALYZE TABLE
BACKUP
BATCH
BEGIN
CHANGE COLUMN
COMMIT
CHANGE DRAINER
CHANGE PUMP
CREATE [GLOBAL|SESSION] BINDING
CREATE DATABASE
CREATE INDEX
CREATE PLACEMENT POLICY
CREATE ROLE
CREATE SEQUENCE
CREATE TABLE LIKE
CREATE TABLE
CREATE USER
CREATE VIEW
DEALLOCATE
DELETE
DESC
DESCRIBE
DO
DROP [GLOBAL|SESSION] BINDING
DROP COLUMN
DROP DATABASE
DROP INDEX
DROP PLACEMENT POLICY
DROP ROLE
DROP SEQUENCE
DROP STATS
DROP TABLE
DROP USER
DROP VIEW
EXECUTE
EXPLAIN ANALYZE
EXPLAIN
FLASHBACK TABLE
FLUSH PRIVILEGES
FLUSH STATUS
FLUSH TABLES
GRANT <privileges>
GRANT <role>
INSERT
KILL [TIDB]
LOAD DATA
LOAD STATS
MODIFY COLUMN
PREPARE
RECOVER TABLE
RENAME INDEX
RENAME TABLE
REPLACE
RESTORE
REVOKE <privileges>
REVOKE <role>
ROLLBACK
SELECT
SET DEFAULT ROLE
SET [NAMES|CHARACTER SET]
SET PASSWORD
SET ROLE
SET TRANSACTION
SET [GLOBAL|SESSION] <variable>
SHOW ANALYZE STATUS
SHOW [BACKUPS|RESTORES]
SHOW [GLOBAL|SESSION] BINDINGS
SHOW BUILTINS
SHOW CHARACTER SET
SHOW COLLATION
SHOW [FULL] COLUMNS FROM
SHOW CONFIG
SHOW CREATE PLACEMENT POLICY
SHOW CREATE SEQUENCE
SHOW CREATE TABLE
SHOW CREATE USER
SHOW DATABASES
SHOW DRAINER STATUS
SHOW ENGINES
SHOW ERRORS
SHOW [FULL] FIELDS FROM
SHOW GRANTS
SHOW INDEX [FROM|IN]
SHOW INDEXES [FROM|IN]
SHOW KEYS [FROM|IN]
SHOW MASTER STATUS
SHOW PLACEMENT
SHOW PLACEMENT FOR
SHOW PLACEMENT LABELS
SHOW PLUGINS
SHOW PRIVILEGES
SHOW [FULL] PROCESSSLIST
SHOW PROFILES
SHOW PUMP STATUS
SHOW SCHEMAS
SHOW STATS_HEALTHY
SHOW STATS_HISTOGRAMS
SHOW STATS_META
SHOW STATUS
SHOW TABLE NEXT_ROW_ID
SHOW TABLE REGIONS
SHOW TABLE STATUS
SHOW [FULL] TABLES
SHOW [GLOBAL|SESSION] VARIABLES
SHOW WARNINGS
SHUTDOWN
SPLIT REGION
START TRANSACTION
TABLE
TRACE
TRUNCATE
UPDATE
USE
WITH
- Data Types
- Functions and Operators
- Overview
- Type Conversion in Expression Evaluation
- Operators
- Control Flow Functions
- String Functions
- Numeric Functions and Operators
- Date and Time Functions
- Bit Functions and Operators
- Cast Functions and Operators
- Encryption and Compression Functions
- Locking Functions
- Information Functions
- JSON Functions
- Aggregate (GROUP BY) Functions
- Window Functions
- Miscellaneous Functions
- Precision Math
- Set Operations
- List of Expressions for Pushdown
- TiDB Specific Functions
- Clustered Indexes
- Constraints
- Generated Columns
- SQL Mode
- Table Attributes
- Transactions
- Garbage Collection (GC)
- Views
- Partitioning
- Temporary Tables
- Cached Tables
- Character Set and Collation
- Placement Rules in SQL
- System Tables
mysql
- INFORMATION_SCHEMA
- Overview
ANALYZE_STATUS
CLIENT_ERRORS_SUMMARY_BY_HOST
CLIENT_ERRORS_SUMMARY_BY_USER
CLIENT_ERRORS_SUMMARY_GLOBAL
CHARACTER_SETS
CLUSTER_CONFIG
CLUSTER_HARDWARE
CLUSTER_INFO
CLUSTER_LOAD
CLUSTER_LOG
CLUSTER_SYSTEMINFO
COLLATIONS
COLLATION_CHARACTER_SET_APPLICABILITY
COLUMNS
DATA_LOCK_WAITS
DDL_JOBS
DEADLOCKS
ENGINES
INSPECTION_RESULT
INSPECTION_RULES
INSPECTION_SUMMARY
KEY_COLUMN_USAGE
METRICS_SUMMARY
METRICS_TABLES
PARTITIONS
PLACEMENT_POLICIES
PROCESSLIST
REFERENTIAL_CONSTRAINTS
SCHEMATA
SEQUENCES
SESSION_VARIABLES
SLOW_QUERY
STATISTICS
TABLES
TABLE_CONSTRAINTS
TABLE_STORAGE_STATS
TIDB_HOT_REGIONS
TIDB_HOT_REGIONS_HISTORY
TIDB_INDEXES
TIDB_SERVERS_INFO
TIDB_TRX
TIFLASH_REPLICA
TIKV_REGION_PEERS
TIKV_REGION_STATUS
TIKV_STORE_STATUS
USER_PRIVILEGES
VIEWS
METRICS_SCHEMA
- UI
- TiDB Dashboard
- Overview
- Maintain
- Access
- Overview Page
- Cluster Info Page
- Top SQL Page
- Key Visualizer Page
- Metrics Relation Graph
- SQL Statements Analysis
- Slow Queries Page
- Cluster Diagnostics
- Search Logs Page
- Instance Profiling
- Session Management and Configuration
- FAQ
- CLI
- Command Line Flags
- Configuration File Parameters
- System Variables
- Storage Engines
- Telemetry
- Errors Codes
- Table Filter
- Schedule Replicas by Topology Labels
- FAQs
- Release Notes
- All Releases
- Release Timeline
- TiDB Versioning
- v6.1
- v6.0
- v5.4
- v5.3
- v5.2
- v5.1
- v5.0
- v4.0
- v3.1
- v3.0
- v2.1
- v2.0
- v1.0
- Glossary
TiDB Specific Functions
The following functions are TiDB extensions, and are not present in MySQL:
Function name | Function description |
---|---|
TIDB_BOUNDED_STALENESS() | The TIDB_BOUNDED_STALENESS function instructs TiDB to read the data as new as possible within the time range. See also: Read Historical Data Using the AS OF TIMESTAMP Clause |
TIDB_DECODE_KEY(str) | The TIDB_DECODE_KEY function can be used to decode a TiDB-encoded key entry into a JSON structure containing _tidb_rowid and table_id . These encoded keys can be found in some system tables and in logging outputs. |
TIDB_DECODE_PLAN(str) | The TIDB_DECODE_PLAN function can be used to decode a TiDB execution plan. |
TIDB_IS_DDL_OWNER() | The TIDB_IS_DDL_OWNER function can be used to check whether or not the TiDB instance you are connected to is the one that is the DDL Owner. The DDL Owner is the TiDB instance that is tasked with executing DDL statements on behalf of all other nodes in the cluster. |
TIDB_PARSE_TSO(num) | The TIDB_PARSE_TSO function can be used to extract the physical timestamp from a TiDB TSO timestamp. See also: tidb_current_ts . |
TIDB_VERSION() | The TIDB_VERSION function returns the TiDB version with additional build information. |
TIDB_DECODE_SQL_DIGESTS(digests, stmtTruncateLength) | The TIDB_DECODE_SQL_DIGESTS() function is used to query the normalized SQL statements (a form without formats and arguments) corresponding to the set of SQL digests in the cluster. |
VITESS_HASH(str) | The VITESS_HASH function returns the hash of a string that is compatible with Vitess' HASH function. This is intended to help the data migration from Vitess. |
TIDB_SHARD() | The TIDB_SHARD function can be used to create a shard index to scatter the index hotspot. A shard index is an expression index with a TIDB_SHARD function as the prefix. |
Examples
This section provides examples for some of the functions above.
TIDB_DECODE_KEY
In the following example, the table t1
has a hidden rowid
that is generated by TiDB. The TIDB_DECODE_KEY
is used in the statement. From the result, you can see that the hidden rowid
is decoded and output, which is a typical result for the non-clustered primary key.
SELECT START_KEY, TIDB_DECODE_KEY(START_KEY) FROM information_schema.tikv_region_status WHERE table_name='t1' AND REGION_ID=2\G
*************************** 1. row ***************************
START_KEY: 7480000000000000FF3B5F728000000000FF1DE3F10000000000FA
TIDB_DECODE_KEY(START_KEY): {"_tidb_rowid":1958897,"table_id":"59"}
1 row in set (0.00 sec)
In the following example, the table t2
has a compound clustered primary key. From the JSON output, you can see a handle
that contains the name and value for both of the columns that are part of the primary key.
show create table t2\G
*************************** 1. row ***************************
Table: t2
Create Table: CREATE TABLE `t2` (
`id` binary(36) NOT NULL,
`a` tinyint(3) unsigned NOT NULL,
`v` varchar(512) DEFAULT NULL,
PRIMARY KEY (`a`,`id`) /*T![clustered_index] CLUSTERED */
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_bin
1 row in set (0.001 sec)
select * from information_schema.tikv_region_status where table_name='t2' limit 1\G
*************************** 1. row ***************************
REGION_ID: 48
START_KEY: 7480000000000000FF3E5F720400000000FF0000000601633430FF3338646232FF2D64FF3531632D3131FF65FF622D386337352DFFFF3830653635303138FFFF61396265000000FF00FB000000000000F9
END_KEY:
TABLE_ID: 62
DB_NAME: test
TABLE_NAME: t2
IS_INDEX: 0
INDEX_ID: NULL
INDEX_NAME: NULL
EPOCH_CONF_VER: 1
EPOCH_VERSION: 38
WRITTEN_BYTES: 0
READ_BYTES: 0
APPROXIMATE_SIZE: 136
APPROXIMATE_KEYS: 479905
REPLICATIONSTATUS_STATE: NULL
REPLICATIONSTATUS_STATEID: NULL
1 row in set (0.005 sec)
select tidb_decode_key('7480000000000000FF3E5F720400000000FF0000000601633430FF3338646232FF2D64FF3531632D3131FF65FF622D386337352DFFFF3830653635303138FFFF61396265000000FF00FB000000000000F9');
+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| tidb_decode_key('7480000000000000FF3E5F720400000000FF0000000601633430FF3338646232FF2D64FF3531632D3131FF65FF622D386337352DFFFF3830653635303138FFFF61396265000000FF00FB000000000000F9') |
+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| {"handle":{"a":"6","id":"c4038db2-d51c-11eb-8c75-80e65018a9be"},"table_id":62} |
+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1 row in set (0.001 sec)
TIDB_DECODE_PLAN
You can find TiDB execution plans in encoded form in the slow query log. The TIDB_DECODE_PLAN()
function is then used to decode the encoded plans into a human-readable form.
This function is useful because a plan is captured at the time the statement is executed. Re-executing the statement in EXPLAIN
might produce different results as data distribution and statistics evolves over time.
SELECT tidb_decode_plan('8QIYMAkzMV83CQEH8E85LjA0CWRhdGE6U2VsZWN0aW9uXzYJOTYwCXRpbWU6NzEzLjHCtXMsIGxvb3BzOjIsIGNvcF90YXNrOiB7bnVtOiAxLCBtYXg6IDU2OC41wgErRHByb2Nfa2V5czogMCwgcnBjXxEpAQwFWBAgNTQ5LglZyGNvcHJfY2FjaGVfaGl0X3JhdGlvOiAwLjAwfQkzLjk5IEtCCU4vQQoxCTFfNgkxXzAJMwm2SGx0KHRlc3QudC5hLCAxMDAwMCkNuQRrdgmiAHsFbBQzMTMuOMIBmQnEDDk2MH0BUgEEGAoyCTQzXzUFVwX1oGFibGU6dCwga2VlcCBvcmRlcjpmYWxzZSwgc3RhdHM6cHNldWRvCTk2ISE2aAAIMTUzXmYA')\G
*************************** 1. row ***************************
tidb_decode_plan('8QIYMAkzMV83CQEH8E85LjA0CWRhdGE6U2VsZWN0aW9uXzYJOTYwCXRpbWU6NzEzLjHCtXMsIGxvb3BzOjIsIGNvcF90YXNrOiB7bnVtOiAxLCBtYXg6IDU2OC41wgErRHByb2Nfa2V5czogMCwgcnBjXxEpAQwFWBAgNTQ5LglZyGNvcHJfY2FjaGVfaGl0X3JhdGlvOiAwLjAwfQkzLjk5IEtCCU4vQQoxCTFfNgkxXz: id task estRows operator info actRows execution info memory disk
TableReader_7 root 319.04 data:Selection_6 960 time:713.1µs, loops:2, cop_task: {num: 1, max: 568.5µs, proc_keys: 0, rpc_num: 1, rpc_time: 549.1µs, copr_cache_hit_ratio: 0.00} 3.99 KB N/A
└─Selection_6 cop[tikv] 319.04 lt(test.t.a, 10000) 960 tikv_task:{time:313.8µs, loops:960} N/A N/A
└─TableFullScan_5 cop[tikv] 960 table:t, keep order:false, stats:pseudo 960 tikv_task:{time:153µs, loops:960} N/A N/A
TIDB_PARSE_TSO
The TIDB_PARSE_TSO
function can be used to extract the physical timestamp from a TiDB TSO timestamp. TSO stands for Time Stamp Oracle and is a monotonically increasing timestamp given out by PD (Placement Driver) for every transaction.
A TSO is a number that consists of two parts:
- A physical timestamp
- A logical counter
BEGIN;
SELECT TIDB_PARSE_TSO(@@tidb_current_ts);
ROLLBACK;
+-----------------------------------+
| TIDB_PARSE_TSO(@@tidb_current_ts) |
+-----------------------------------+
| 2021-05-26 11:33:37.776000 |
+-----------------------------------+
1 row in set (0.0012 sec)
Here TIDB_PARSE_TSO
is used to extract the physical timestamp from the timestamp number that is available in the tidb_current_ts
session variable. Because timestamps are given out per transaction, this function is running in a transaction.
TIDB_VERSION
The TIDB_VERSION
function can be used to get the version and build details of the TiDB server that you are connected to. You can use this function when reporting issues on GitHub.
SELECT TIDB_VERSION()\G
*************************** 1. row ***************************
TIDB_VERSION(): Release Version: v5.1.0-alpha-13-gd5e0ed0aa-dirty
Edition: Community
Git Commit Hash: d5e0ed0aaed72d2f2dfe24e9deec31cb6cb5fdf0
Git Branch: master
UTC Build Time: 2021-05-24 14:39:20
GoVersion: go1.13
Race Enabled: false
TiKV Min Version: v3.0.0-60965b006877ca7234adaced7890d7b029ed1306
Check Table Before Drop: false
1 row in set (0.00 sec)
TIDB_DECODE_SQL_DIGESTS
The TIDB_DECODE_SQL_DIGESTS()
function is used to query the normalized SQL statements (a form without formats and arguments) corresponding to the set of SQL digests in the cluster. This function accepts 1 or 2 arguments:
digests
: A string. This parameter is in the format of a JSON string array, and each string in the array is a SQL digest.stmtTruncateLength
: An integer (optional). It is used to limit the length of each SQL statement in the returned result. If a SQL statement exceeds the specified length, the statement is truncated.0
means that the length is unlimited.
This function returns a string, which is in the format of a JSON string array. The i-th item in the array is the normalized SQL statement corresponding to the i-th element in the digests
parameter. If an element in the digests
parameter is not a valid SQL digest or the system cannot find the corresponding SQL statement, the corresponding item in the returned result is null
. If the truncation length is specified (stmtTruncateLength > 0
), for each statement in the returned result that exceeds this length, the first stmtTruncateLength
characters are retained and the suffix "..."
is added at the end to indicate the truncation. If the digests
parameter is NULL
, the returned value of the function is NULL
.
- Only users with the PROCESS privilege can use this function.
- When
TIDB_DECODE_SQL_DIGESTS
is executed, TiDB queries the statement corresponding to each SQL digest from the statement summary tables, so there is no guarantee that the corresponding statement can always be found for any SQL digest. Only the statements that have been executed in the cluster can be found, and whether these SQL statements can be queried or not is also affected by the related configuration of the statement summary tables. For the detailed description of the statement summary table, see Statement Summary Tables. - This function has a high overhead. In queries with a large number of rows (for example, querying the full table of
information_schema.cluster_tidb_trx
on a large and busy cluster), using this function might cause the queries to run for too long. Use it with caution.- This function has a high overhead because every time it is called, it internally queries the
STATEMENTS_SUMMARY
,STATEMENTS_SUMMARY_HISTORY
,CLUSTER_STATEMENTS_SUMMARY
, andCLUSTER_STATEMENTS_SUMMARY_HISTORY
tables, and the query involves theUNION
operation. This function currently does not support vectorization, that is, when calling this function for multiple rows of data, the above query is performed separately for each row.
- This function has a high overhead because every time it is called, it internally queries the
set @digests = '["e6f07d43b5c21db0fbb9a31feac2dc599787763393dd5acbfad80e247eb02ad5","38b03afa5debbdf0326a014dbe5012a62c51957f1982b3093e748460f8b00821","e5796985ccafe2f71126ed6c0ac939ffa015a8c0744a24b7aee6d587103fd2f7"]';
select tidb_decode_sql_digests(@digests);
+------------------------------------+
| tidb_decode_sql_digests(@digests) |
+------------------------------------+
| ["begin",null,"select * from `t`"] |
+------------------------------------+
1 row in set (0.00 sec)
In the above example, the parameter is a JSON array containing 3 SQL digests, and the corresponding SQL statements are the three items in the query results. But the SQL statement corresponding to the second SQL digest cannot be found from the cluster, so the second item in the result is null
.
select tidb_decode_sql_digests(@digests, 10);
+---------------------------------------+
| tidb_decode_sql_digests(@digests, 10) |
+---------------------------------------+
| ["begin",null,"select * f..."] |
+---------------------------------------+
1 row in set (0.01 sec)
The above call specifies the second parameter (that is, the truncation length) as 10, and the length of the third statement in the query result is greater than 10. Therefore, only the first 10 characters are retained, and "..."
is added at the end, which indicates the truncation.
See also:
TIDB_SHARD
The TIDB_SHARD
function can be used to create a shard index to scatter the index hotspot. A shard index is an expression index prefixed with a TIDB_SHARD
function.
Shard index
Creation:
To create a shard index for the index field
a
, you can useuk((tidb_shard(a)), a))
. When there is a hotspot caused by monotonically increasing or decreasing data on the index fielda
in the unique secondary indexuk((tidb_shard(a)), a))
, the index's prefixtidb_shard(a)
can scatter the hotspot to improve the scalability of the cluster.Scenarios:
- There is a write hotspot caused by monotonically increasing or decreasing keys on the unique secondary index, and the index contains integer type fields.
- The SQL statement executes an equality query based on all fields of the secondary index, either as a separate
SELECT
or as an internal query generated byUPDATE
,DELETE
and so on. The equality query includes two ways:a = 1
ora IN (1, 2, ......)
.
Limitations:
- Cannot be used in inequality queries.
- Cannot be used in queries that contain
OR
mixed with an outmostAND
operator. - Cannot be used in the
GROUP BY
clause. - Cannot be used in the
ORDER BY
clause. - Cannot be used in the
ON
clause. - Cannot be used in the
WHERE
subquery. - Can be used to scatter unique indexes of only the integer fields.
- Might not take effect in composite indexes.
- Cannot go through FastPlan process, which affects optimizer performance.
- Cannot be used to prepare the execution plan cache.
Synopsis
- TIDBShardExpr
TIDBShardExpr ::=
"TIDB_SHARD" "(" expr ")"
Example
Use the
TIDB_SHARD
function to calculate the SHARD value.The following statement shows how to use the
TIDB_SHARD
function to calculate the SHARD value of12373743746
:SELECT TIDB_SHARD(12373743746);
The SHARD value is:
+-------------------------+ | TIDB_SHARD(12373743746) | +-------------------------+ | 184 | +-------------------------+ 1 row in set (0.00 sec)
Create a shard index using the
TIDB_SHARD
function:CREATE TABLE test(id INT PRIMARY KEY CLUSTERED, a INT, b INT, UNIQUE KEY uk((tidb_shard(a)), a));