sql-formatter-org · nene · Mar 16, 2024 · Mar 11, 2024 · Mar 11, 2024 · Mar 11, 2024
diff --git a/AUTHORS b/AUTHORS
@@ -13,6 +13,7 @@ Boris Verkhovskiy <boris.verk@gmail.com>
 Christian Jorgensen <chr.jorgensen1@gmail.com>
 Christopher Manouvrier <chris@dovetailapp.com>
 Damon Davison <ddavison@avalere.com>
+Daniël van Eeden <daniel.van.eeden@pingcap.com>
 Davut Can Abacigil <can@teamsql.io>
 eeeXun <sdes96303@gmail.com>
 Erik Hirmo <erik.hirmo@roguewave.com>

diff --git a/README.md b/README.md
@@ -7,7 +7,7 @@
 It started as a port of a [PHP Library][], but has since considerably diverged.
 
 It supports various SQL dialects:
-GCP BigQuery, IBM DB2, Apache Hive, MariaDB, MySQL, Couchbase N1QL, Oracle PL/SQL, PostgreSQL, Amazon Redshift, SingleStoreDB, Snowflake, Spark, SQL Server Transact-SQL, Trino (and Presto).
+GCP BigQuery, IBM DB2, Apache Hive, MariaDB, MySQL, TiDB, Couchbase N1QL, Oracle PL/SQL, PostgreSQL, Amazon Redshift, SingleStoreDB, Snowflake, Spark, SQL Server Transact-SQL, Trino (and Presto).
 See [language option docs](docs/language.md) for more details.
 
 It does not support:
@@ -123,7 +123,7 @@ sql-formatter -h
 
 ```
 usage: sql-formatter [-h] [-o OUTPUT] \
-[-l {bigquery,db2,db2i,hive,mariadb,mysql,n1ql,plsql,postgresql,redshift,singlestoredb,snowflake,spark,sql,sqlite,transactsql,trino,tsql}] [-c CONFIG] [--version] [FILE]
+[-l {bigquery,db2,db2i,hive,mariadb,mysql,n1ql,plsql,postgresql,redshift,singlestoredb,snowflake,spark,sql,sqlite,tidb,transactsql,trino,tsql}] [-c CONFIG] [--version] [FILE]
 
 SQL Formatter
 
@@ -135,7 +135,7 @@ optional arguments:
   -o, --output    OUTPUT
                     File to write SQL output (defaults to stdout)
   --fix           Update the file in-place
-  -l, --language  {bigquery,db2,db2i,hive,mariadb,mysql,n1ql,plsql,postgresql,redshift,singlestoredb,snowflake,spark,sql,sqlite,trino,tsql}
+  -l, --language  {bigquery,db2,db2i,hive,mariadb,mysql,n1ql,plsql,postgresql,redshift,singlestoredb,snowflake,spark,sql,sqlite,tidb,trino,tsql}
                     SQL dialect (defaults to basic sql)
   -c, --config    CONFIG
                     Path to config JSON file or json string (will use default configs if unspecified)

diff --git a/docs/dialect.md b/docs/dialect.md
@@ -27,6 +27,7 @@ The following dialects can be imported from `"sql-formatter"` module:
 - `hive` - [Apache Hive][]
 - `mariadb` - [MariaDB][]
 - `mysql` - [MySQL][]
+- `tidb` - [TiDB][]
 - `n1ql` - [Couchbase N1QL][]
 - `plsql` - [Oracle PL/SQL][]
 - `postgresql` - [PostgreSQL][]
@@ -75,6 +76,7 @@ You likely only want to use this if your other alternative is to fork SQL Format
 [apache hive]: https://hive.apache.org/
 [mariadb]: https://mariadb.com/
 [mysql]: https://www.mysql.com/
+[tidb]: https://github.com/pingcap/tidb/
 [couchbase n1ql]: http://www.couchbase.com/n1ql
 [oracle pl/sql]: http://www.oracle.com/technetwork/database/features/plsql/index.html
 [postgresql]: https://www.postgresql.org/

diff --git a/docs/language.md b/docs/language.md
@@ -19,6 +19,7 @@ const result = format('SELECT * FROM tbl', { dialect: 'sqlite' });
 - `"hive"` - [Apache Hive][]
 - `"mariadb"` - [MariaDB][]
 - `"mysql"` - [MySQL][]
+- `"tidb"` - [TiDB][]
 - `"n1ql"` - [Couchbase N1QL][]
 - `"plsql"` - [Oracle PL/SQL][]
 - `"postgresql"` - [PostgreSQL][]
@@ -52,6 +53,7 @@ See docs for [dialect][] option.
 [apache hive]: https://hive.apache.org/
 [mariadb]: https://mariadb.com/
 [mysql]: https://www.mysql.com/
+[tidb]: https://github.com/pingcap/tidb/
 [couchbase n1ql]: http://www.couchbase.com/n1ql
 [oracle pl/sql]: http://www.oracle.com/technetwork/database/features/plsql/index.html
 [postgresql]: https://www.postgresql.org/

diff --git a/docs/params.md b/docs/params.md
@@ -121,6 +121,7 @@ The placeholder types available by default depend on SQL dialect used:
 - snowflake - _no support_
 - sqlite - `?`, `?1`, `:name`, `@name`, `$name`
 - spark - _no support_
+- tidb - `?`
 - tsql - `@name`, `@"name"`, `@[name]`
 - trino - _no support_
 

diff --git a/package.json b/package.json
@@ -35,6 +35,7 @@
     "db2",
     "db2i",
     "sqlite",
+    "tidb",
     "trino",
     "presto",
     "prestosql",

diff --git a/src/allDialects.ts b/src/allDialects.ts
@@ -4,6 +4,7 @@ export { db2i } from './languages/db2i/db2i.formatter.js';
 export { hive } from './languages/hive/hive.formatter.js';
 export { mariadb } from './languages/mariadb/mariadb.formatter.js';
 export { mysql } from './languages/mysql/mysql.formatter.js';
+export { tidb } from './languages/tidb/tidb.formatter.js';
 export { n1ql } from './languages/n1ql/n1ql.formatter.js';
 export { plsql } from './languages/plsql/plsql.formatter.js';
 export { postgresql } from './languages/postgresql/postgresql.formatter.js';

diff --git a/src/index.ts b/src/index.ts
@@ -9,6 +9,7 @@ export { db2i } from './languages/db2i/db2i.formatter.js';
 export { hive } from './languages/hive/hive.formatter.js';
 export { mariadb } from './languages/mariadb/mariadb.formatter.js';
 export { mysql } from './languages/mysql/mysql.formatter.js';
+export { tidb } from './languages/tidb/tidb.formatter.js';
 export { n1ql } from './languages/n1ql/n1ql.formatter.js';
 export { plsql } from './languages/plsql/plsql.formatter.js';
 export { postgresql } from './languages/postgresql/postgresql.formatter.js';

diff --git a/src/languages/tidb/tidb.formatter.ts b/src/languages/tidb/tidb.formatter.ts
@@ -0,0 +1,221 @@
+import { DialectOptions } from '../../dialect.js';
+import { expandPhrases } from '../../expandPhrases.js';
+import { postProcess } from '../mariadb/likeMariaDb.js';
+import { dataTypes, keywords } from './tidb.keywords.js';
+import { functions } from './tidb.functions.js';
+
+const reservedSelect = expandPhrases(['SELECT [ALL | DISTINCT | DISTINCTROW]']);
+
+const reservedClauses = expandPhrases([
+  // queries
+  'WITH [RECURSIVE]',
+  'FROM',
+  'WHERE',
+  'GROUP BY',
+  'HAVING',
+  'WINDOW',
+  'PARTITION BY',
+  'ORDER BY',
+  'LIMIT',
+  'OFFSET',
+  // Data manipulation
+  // - insert:
+  'INSERT [LOW_PRIORITY | DELAYED | HIGH_PRIORITY] [IGNORE] [INTO]',
+  'REPLACE [LOW_PRIORITY | DELAYED] [INTO]',
+  'VALUES',
+  'ON DUPLICATE KEY UPDATE',
+  // - update:
+  'SET',
+]);
+
+const standardOnelineClauses = expandPhrases(['CREATE [TEMPORARY] TABLE [IF NOT EXISTS]']);
+
+const tabularOnelineClauses = expandPhrases([
+  // https://docs.pingcap.com/tidb/stable/sql-statement-create-view
+  'CREATE [OR REPLACE] [SQL SECURITY DEFINER | SQL SECURITY INVOKER] VIEW [IF NOT EXISTS]',
+  // https://docs.pingcap.com/tidb/stable/sql-statement-update
+  'UPDATE [LOW_PRIORITY] [IGNORE]',
+  // https://docs.pingcap.com/tidb/stable/sql-statement-delete
+  'DELETE [LOW_PRIORITY] [QUICK] [IGNORE] FROM',
+  // https://docs.pingcap.com/tidb/stable/sql-statement-drop-table
+  'DROP [TEMPORARY] TABLE [IF EXISTS]',
+  // https://docs.pingcap.com/tidb/stable/sql-statement-alter-table
+  'ALTER TABLE',
+  'ADD [COLUMN]',
+  '{CHANGE | MODIFY} [COLUMN]',
+  'DROP [COLUMN]',
+  'RENAME [TO | AS]',
+  'RENAME COLUMN',
+  'ALTER [COLUMN]',
+  '{SET | DROP} DEFAULT', // for alter column
+  // https://docs.pingcap.com/tidb/stable/sql-statement-truncate
+  'TRUNCATE [TABLE]',
+  // https://docs.pingcap.com/tidb/stable/sql-statement-alter-database
+  'ALTER DATABASE',
+  // https://docs.pingcap.com/tidb/stable/sql-statement-alter-instance
+  'ALTER INSTANCE',
+  'ALTER RESOURCE GROUP',
+  'ALTER SEQUENCE',
+  // https://docs.pingcap.com/tidb/stable/sql-statement-alter-user
+  'ALTER USER',
+  'ALTER VIEW',
+  'ANALYZE TABLE',
+  'CHECK TABLE',
+  'CHECKSUM TABLE',
+  'COMMIT',
+  'CREATE DATABASE',
+  'CREATE INDEX',
+  'CREATE RESOURCE GROUP',
+  'CREATE ROLE',
+  'CREATE SEQUENCE',
+  'CREATE USER',
+  'DEALLOCATE PREPARE',
+  'DESCRIBE',
+  'DROP DATABASE',
+  'DROP INDEX',
+  'DROP RESOURCE GROUP',
+  'DROP ROLE',
+  'DROP TABLESPACE',
+  'DROP USER',
+  'DROP VIEW',
+  'EXPLAIN',
+  'FLUSH',
+  // https://docs.pingcap.com/tidb/stable/sql-statement-grant-privileges
+  'GRANT',
+  'IMPORT TABLE',
+  'INSTALL COMPONENT',
+  'INSTALL PLUGIN',
+  'KILL',
+  'LOAD DATA',
+  'LOCK INSTANCE FOR BACKUP',
+  'LOCK TABLES',
+  'OPTIMIZE TABLE',
+  'PREPARE',
+  'RELEASE SAVEPOINT',
+  'RENAME TABLE',
+  'RENAME USER',
+  'REPAIR TABLE',
+  'RESET',
+  'REVOKE',
+  'ROLLBACK',
+  'ROLLBACK TO SAVEPOINT',
+  'SAVEPOINT',
+  'SET CHARACTER SET',
+  'SET DEFAULT ROLE',
+  'SET NAMES',
+  'SET PASSWORD',
+  'SET RESOURCE GROUP',
+  'SET ROLE',
+  'SET TRANSACTION',
+  'SHOW',
+  'SHOW BINARY LOGS',
+  'SHOW BINLOG EVENTS',
+  'SHOW CHARACTER SET',
+  'SHOW COLLATION',
+  'SHOW COLUMNS',
+  'SHOW CREATE DATABASE',
+  'SHOW CREATE TABLE',
+  'SHOW CREATE USER',
+  'SHOW CREATE VIEW',
+  'SHOW DATABASES',
+  'SHOW ENGINE',
+  'SHOW ENGINES',
+  'SHOW ERRORS',
+  'SHOW EVENTS',
+  'SHOW GRANTS',
+  'SHOW INDEX',
+  'SHOW MASTER STATUS',
+  'SHOW OPEN TABLES',
+  'SHOW PLUGINS',
+  'SHOW PRIVILEGES',
+  'SHOW PROCESSLIST',
+  'SHOW PROFILE',
+  'SHOW PROFILES',
+  'SHOW STATUS',
+  'SHOW TABLE STATUS',
+  'SHOW TABLES',
+  'SHOW TRIGGERS',
+  'SHOW VARIABLES',
+  'SHOW WARNINGS',
+  // https://docs.pingcap.com/tidb/stable/sql-statement-table
+  'TABLE',
+  'UNINSTALL COMPONENT',
+  'UNINSTALL PLUGIN',
+  'UNLOCK INSTANCE',
+  'UNLOCK TABLES',
+  // https://docs.pingcap.com/tidb/stable/sql-statement-use
+  'USE',
+]);
+
+const reservedSetOperations = expandPhrases(['UNION [ALL | DISTINCT]']);
+
+const reservedJoins = expandPhrases([
+  'JOIN',
+  '{LEFT | RIGHT} [OUTER] JOIN',
+  '{INNER | CROSS} JOIN',
+  'NATURAL [INNER] JOIN',
+  'NATURAL {LEFT | RIGHT} [OUTER] JOIN',
+  // non-standard joins
+  'STRAIGHT_JOIN',
+]);
+
+const reservedPhrases = expandPhrases([
+  'ON {UPDATE | DELETE} [SET NULL]',
+  'CHARACTER SET',
+  '{ROWS | RANGE} BETWEEN',
+  'IDENTIFIED BY',
+]);
+
+// https://docs.pingcap.com/tidb/stable/basic-features
+export const tidb: DialectOptions = {
+  name: 'tidb',
+  tokenizerOptions: {
+    reservedSelect,
+    reservedClauses: [...reservedClauses, ...standardOnelineClauses, ...tabularOnelineClauses],
+    reservedSetOperations,
+    reservedJoins,
+    reservedPhrases,
+    supportsXor: true,
+    reservedKeywords: keywords,
+    reservedDataTypes: dataTypes,
+    reservedFunctionNames: functions,
+    // TODO: support _ char set prefixes such as _utf8, _latin1, _binary, _utf8mb4, etc.
+    stringTypes: [
+      '""-qq-bs',
+      { quote: "''-qq-bs", prefixes: ['N'] },
+      { quote: "''-raw", prefixes: ['B', 'X'], requirePrefix: true },
+    ],
+    identTypes: ['``'],
+    identChars: { first: '$', rest: '$', allowFirstCharNumber: true },
+    variableTypes: [
+      { regex: '@@?[A-Za-z0-9_.$]+' },
+      { quote: '""-qq-bs', prefixes: ['@'], requirePrefix: true },
+      { quote: "''-qq-bs", prefixes: ['@'], requirePrefix: true },
+      { quote: '``', prefixes: ['@'], requirePrefix: true },
+    ],
+    paramTypes: { positional: true },
+    lineCommentTypes: ['--', '#'],
+    operators: [
+      '%',
+      ':=',
+      '&',
+      '|',
+      '^',
+      '~',
+      '<<',
+      '>>',
+      '<=>',
+      '->',
+      '->>',
+      '&&',
+      '||',
+      '!',
+      '*.*', // Not actually an operator
+    ],
+    postProcess,
+  },
+  formatOptions: {
+    onelineClauses: [...standardOnelineClauses, ...tabularOnelineClauses],
+    tabularOnelineClauses,
+  },
+};