Commands

Index

Commands (Data Manipulation)

SELECT INSERT UPDATE DELETE BACKUP

CALL EXPLAIN MERGE MERGE USING RUNSCRIPT

SCRIPT SHOW Explicit table Table value WITH

Commands (Data Definition)

ALTER INDEX RENAME ALTER SCHEMA RENAME ALTER SEQUENCE ALTER TABLE ADD ALTER TABLE ADD CONSTRAINT ALTER TABLE RENAME CONSTRAINT ALTER TABLE ALTER COLUMN ALTER TABLE DROP COLUMN ALTER TABLE DROP CONSTRAINT ALTER TABLE SET ALTER TABLE RENAME ALTER USER ADMIN ALTER USER RENAME ALTER USER SET PASSWORD ALTER VIEW RECOMPILE

ALTER VIEW RENAME ANALYZE COMMENT CREATE AGGREGATE CREATE ALIAS CREATE CONSTANT CREATE DOMAIN CREATE INDEX CREATE LINKED TABLE CREATE ROLE CREATE SCHEMA CREATE SEQUENCE CREATE TABLE CREATE TRIGGER CREATE USER

CREATE VIEW DROP AGGREGATE DROP ALIAS DROP ALL OBJECTS DROP CONSTANT DROP DOMAIN DROP INDEX DROP ROLE DROP SCHEMA DROP SEQUENCE DROP TABLE DROP TRIGGER DROP USER DROP VIEW TRUNCATE TABLE

Commands (Other)

CHECKPOINT CHECKPOINT SYNC COMMIT COMMIT TRANSACTION GRANT RIGHT GRANT ALTER ANY SCHEMA GRANT ROLE HELP PREPARE COMMIT REVOKE RIGHT REVOKE ROLE ROLLBACK ROLLBACK TRANSACTION SAVEPOINT SET @ SET ALLOW_LITERALS SET AUTOCOMMIT SET CACHE_SIZE SET CLUSTER SET BINARY_COLLATION

SET UUID_COLLATION SET BUILTIN_ALIAS_OVERRIDE SET COLLATION SET COMPRESS_LOB SET DATABASE_EVENT_LISTENER SET DB_CLOSE_DELAY SET DEFAULT_LOCK_TIMEOUT SET DEFAULT_TABLE_TYPE SET EXCLUSIVE SET IGNORECASE SET JAVA_OBJECT_SERIALIZER SET LAZY_QUERY_EXECUTION SET LOG SET LOCK_MODE SET LOCK_TIMEOUT SET MAX_LENGTH_INPLACE_LOB SET MAX_LOG_SIZE SET MAX_MEMORY_ROWS SET MAX_MEMORY_UNDO SET MAX_OPERATION_MEMORY

SET MODE SET MULTI_THREADED SET OPTIMIZE_REUSE_RESULTS SET PASSWORD SET QUERY_STATISTICS SET QUERY_STATISTICS_MAX_ENTRIES SET QUERY_TIMEOUT SET REFERENTIAL_INTEGRITY SET RETENTION_TIME SET SALT HASH SET SCHEMA SET SCHEMA_SEARCH_PATH SET THROTTLE SET TRACE_LEVEL SET TRACE_MAX_FILE_SIZE SET UNDO_LOG SET WRITE_DELAY SHUTDOWN

Details

Click on the header to switch between railroad diagram and BNF.

Commands (Data Manipulation)

SELECT

SELECT [ TOP term [ PERCENT ] [ WITH TIES ] ]
[ DISTINCT [ ON ( expression [,...] ) ] | ALL ]
selectExpression [,...]
[ FROM tableExpression [,...] ]
[ WHERE expression ]
[ GROUP BY expression [,...] ] [ HAVING expression ]
[ WINDOW { { windowName AS windowSpecification } [,...] } ]
[ QUALIFY expression ]
[ { UNION [ ALL ] | EXCEPT | MINUS | INTERSECT } select ]
[ ORDER BY order [,...] ]
[ LIMIT expression [ OFFSET expression ] [ SAMPLE_SIZE rowCountInt ] ]
[ [ OFFSET expression { ROW | ROWS } ]
[ FETCH { FIRST | NEXT } [ expression [ PERCENT ] ] { ROW | ROWS }
    { ONLY | WITH TIES } ] [ SAMPLE_SIZE rowCountInt ] ]
[ FOR UPDATE ]

Selects data from a table or multiple tables.

Command is executed in the following logical order:

1. Data is taken from table value expressions that are specified in the FROM clause, joins are executed. If FROM clause is not specified a single row is constructed.

2. WHERE filters rows. Aggregate or window functions are not allowed in this clause.

3. GROUP BY groups the result by the given expression(s). If GROUP BY clause is not specified, but non-window aggregate functions are used or HAVING is specified all rows are grouped together.

4. Aggregate functions are evaluated, SAMPLE_SIZE limits the number of rows read.

5. HAVING filters rows after grouping and evaluation of aggregate functions. Non-window aggregate functions are allowed in this clause.

6. Window functions are evaluated.

7. QUALIFY filters rows after evaluation of window functions. Aggregate and window functions are allowed in this clause.

8. DISTINCT removes duplicates. If DISTINCT ON is used only the specified expressions are checked for duplicates; ORDER BY clause, if any, is used to determine preserved rows. First row is each DISTINCT ON group is preserved. In absence of ORDER BY preserved rows are not determined, database may choose any row from each DISTINCT ON group.

9. UNION, EXCEPT (MINUS), and INTERSECT combine the result of this query with the results of another query. Multiple set operators (UNION, INTERSECT, MINUS, EXCEPT) are evaluated from left to right. For compatibility with other databases and future versions of H2 please use parentheses.

10. ORDER BY sorts the result by the given column(s) or expression(s).

11. Number of rows in output can be limited either with standard OFFSET / FETCH, with non-standard LIMIT / OFFSET, or with non-standard TOP clauses. Different clauses cannot be used together. OFFSET specifies how many rows to skip. Please note that queries with high offset values can be slow. FETCH FIRST/NEXT, LIMIT or TOP limits the number of rows returned by the query (no limit if null or smaller than zero). If PERCENT is specified number of rows is specified as a percent of the total number of rows and should be an integer value between 0 and 100 inclusive. WITH TIES can be used only together with ORDER BY and means that all additional rows that have the same sorting position as the last row will be also returned.

WINDOW clause specifies window definitions for window functions and window aggregate functions. This clause can be used to reuse the same definition in multiple functions.

If FOR UPDATE is specified, the tables or rows are locked for writing. This clause is not allowed in DISTINCT queries and in queries with non-window aggregates, GROUP BY, or HAVING clauses. When using default MVStore engine only the selected rows are locked as in an UPDATE statement. Rows from the right side of a left join and from the left side of a right join, including nested joins, aren't locked. Locking behavior for rows that were excluded from result using OFFSET / FETCH / LIMIT / TOP or QUALIFY is undefined, to avoid possible locking of excessive rows try to filter out unneeded rows with the WHERE criteria when possible. Rows are processed one by one. Committed row is read, tested with WHERE criteria, locked, read again and re-tested, because its value may be changed by concurrent transaction before lock acquisition. Note that new uncommitted rows from other transactions are not visible due to read committed isolation level and therefore cannot be locked. With PageStore engine the whole tables are locked; to avoid deadlocks with this engine always lock the tables in the same order in all transactions.

Example:

SELECT * FROM TEST;
SELECT * FROM TEST ORDER BY NAME;
SELECT ID, COUNT(*) FROM TEST GROUP BY ID;
SELECT NAME, COUNT(*) FROM TEST GROUP BY NAME HAVING COUNT(*) > 2;
SELECT 'ID' COL, MAX(ID) AS MAX FROM TEST UNION SELECT 'NAME', MAX(NAME) FROM TEST;
SELECT * FROM TEST OFFSET 1000 ROWS FETCH FIRST 1000 ROWS ONLY;
SELECT A, B FROM TEST ORDER BY A FETCH FIRST 10 ROWS WITH TIES;
SELECT * FROM (SELECT ID, COUNT(*) FROM TEST
    GROUP BY ID UNION SELECT NULL, COUNT(*) FROM TEST)
    ORDER BY 1 NULLS LAST;
SELECT DISTINCT C1, C2 FROM TEST;
SELECT DISTINCT ON(C1) C1, C2 FROM TEST ORDER BY C1;

INSERT

INSERT INTO tableName insertColumnsAndSource

Inserts a new row / new rows into a table.

When using DIRECT, then the results from the query are directly applied in the target table without any intermediate step.

When using SORTED, b-tree pages are split at the insertion point. This can improve performance and reduce disk usage.

Example:

INSERT INTO TEST VALUES(1, 'Hello')

UPDATE

UPDATE tableName [ [ AS ] newTableAlias ] SET setClauseList
[ WHERE expression ] [ ORDER BY order [,...] ] [ LIMIT expression ]

Updates data in a table. ORDER BY is supported for MySQL compatibility, but it is ignored.

Example:

UPDATE TEST SET NAME='Hi' WHERE ID=1;
UPDATE PERSON P SET NAME=(SELECT A.NAME FROM ADDRESS A WHERE A.ID=P.ID);

DELETE

DELETE [ TOP term ] FROM tableName [ WHERE expression ] [ LIMIT term ]

Deletes rows form a table. If TOP or LIMIT is specified, at most the specified number of rows are deleted (no limit if null or smaller than zero).

Example:

DELETE FROM TEST WHERE ID=2

BACKUP

BACKUP TO fileNameString

Backs up the database files to a .zip file. Objects are not locked, but the backup is transactionally consistent because the transaction log is also copied. Admin rights are required to execute this command.

Example:

BACKUP TO 'backup.zip'

CALL

CALL expression

Calculates a simple expression. This statement returns a result set with one row, except if the called function returns a result set itself. If the called function returns an array, then each element in this array is returned as a column.

Example:

CALL 15*25

EXPLAIN

EXPLAIN { [ PLAN FOR ] | ANALYZE }
{ select | insert | update | delete | merge }

Shows the execution plan for a statement. When using EXPLAIN ANALYZE, the statement is actually executed, and the query plan will include the actual row scan count for each table.

Example:

EXPLAIN SELECT * FROM TEST WHERE ID=1

MERGE

MERGE INTO tableName [ ( columnName [,...] ) ]
[ KEY ( columnName [,...] ) ]
{ insertValues | select }

Updates existing rows, and insert rows that don't exist. If no key column is specified, the primary key columns are used to find the row. If more than one row per new row is affected, an exception is thrown.

Example:

MERGE INTO TEST KEY(ID) VALUES(2, 'World')

MERGE USING

MERGE INTO targetTableName [ [AS] targetAlias]
USING { ( select ) | sourceTableName }[ [AS] sourceAlias ]
ON expression
mergeWhenClause [,...]

Updates or deletes existing rows, and insert rows that don't exist.

The ON clause specifies the matching column expression. Different rows from a source table may not match with the same target row, but one source row may be matched with multiple target rows.

If statement doesn't need a source table a DUAL table can be substituted.

Example:

MERGE INTO TARGET_TABLE AS T USING SOURCE_TABLE AS S
    ON T.ID = S.ID
    WHEN MATCHED AND T.COL2 <> 'FINAL' THEN
        UPDATE SET T.COL1 = S.COL1
    WHEN MATCHED AND T.COL2 = 'FINAL' THEN
        DELETE
    WHEN NOT MATCHED THEN
        INSERT (ID, COL1, COL2) VALUES(S.ID, S.COL1, S.COL2)
MERGE INTO TARGET_TABLE AS T USING (SELECT * FROM SOURCE_TABLE) AS S
    ON T.ID = S.ID
    WHEN MATCHED AND T.COL2 <> 'FINAL' THEN
        UPDATE SET T.COL1 = S.COL1
    WHEN MATCHED AND T.COL2 = 'FINAL' THEN
        DELETE
    WHEN NOT MATCHED THEN
        INSERT VALUES (S.ID, S.COL1, S.COL2)
MERGE INTO TARGET_TABLE USING DUAL ON ID = 1
    WHEN NOT MATCHED THEN INSERT VALUES (1, 'Test')
    WHEN MATCHED THEN UPDATE SET NAME = 'Test'

RUNSCRIPT

RUNSCRIPT FROM fileNameString scriptCompressionEncryption
[ CHARSET charsetString ]

Runs a SQL script from a file. The script is a text file containing SQL statements; each statement must end with ';'. This command can be used to restore a database from a backup. The password must be in single quotes; it is case sensitive and can contain spaces.

Instead of a file name, an URL may be used. To read a stream from the classpath, use the prefix 'classpath:'. See the Pluggable File System section on the Advanced page.

The compression algorithm must match the one used when creating the script. Instead of a file, an URL may be used.

Admin rights are required to execute this command.

Example:

RUNSCRIPT FROM 'backup.sql'
RUNSCRIPT FROM 'classpath:/com/acme/test.sql'

SCRIPT

SCRIPT { [ NODATA ] | [ SIMPLE ] [ COLUMNS ] }
[ NOPASSWORDS ] [ NOSETTINGS ]
[ DROP ] [ BLOCKSIZE blockSizeInt ]
[ TO fileNameString scriptCompressionEncryption
[ CHARSET charsetString ] ]
[ TABLE tableName [, ...] ]
[ SCHEMA schemaName [, ...] ]

Creates a SQL script from the database.

NODATA will not emit INSERT statements. SIMPLE does not use multi-row insert statements. COLUMNS includes column name lists into insert statements. If the DROP option is specified, drop statements are created for tables, views, and sequences. If the block size is set, CLOB and BLOB values larger than this size are split into separate blocks. BLOCKSIZE is used when writing out LOB data, and specifies the point at the values transition from being inserted as inline values, to be inserted using out-of-line commands. NOSETTINGS turns off dumping the database settings (the SET XXX commands)

If no 'TO fileName' clause is specified, the script is returned as a result set. This command can be used to create a backup of the database. For long term storage, it is more portable than copying the database files.

If a 'TO fileName' clause is specified, then the whole script (including insert statements) is written to this file, and a result set without the insert statements is returned.

The password must be in single quotes; it is case sensitive and can contain spaces.

This command locks objects while it is running. Admin rights are required to execute this command.

When using the TABLE or SCHEMA option, only the selected table(s) / schema(s) are included.

Example:

SCRIPT NODATA

SHOW

SHOW { SCHEMAS | TABLES [ FROM schemaName ] |
COLUMNS FROM tableName [ FROM schemaName ] }

Lists the schemas, tables, or the columns of a table.

Example:

SHOW TABLES

Explicit table

TABLE [schemaName.]tableName
[ ORDER BY order [,...] ]
[ OFFSET expression { ROW | ROWS } ]
[ FETCH { FIRST | NEXT } [ expression [ PERCENT ] ] { ROW | ROWS }
{ ONLY | WITH TIES } ]

Selects data from a table.

This command is an equivalent to SELECT * FROM tableName. See SELECT command for description of ORDER BY, OFFSET, and FETCH.

Example:

TABLE TEST;
TABLE TEST ORDER BY ID FETCH FIRST ROW ONLY;

Table value

VALUES rowValueExpression [,...]
[ ORDER BY order [,...] ]
[ OFFSET expression { ROW | ROWS } ]
[ FETCH { FIRST | NEXT } [ expression [ PERCENT ] ] { ROW | ROWS }
{ ONLY | WITH TIES } ]

A list of rows that can be used like a table. See SELECT command for description of ORDER BY, OFFSET, and FETCH. The column list of the resulting table is C1, C2, and so on.

Example:

VALUES (1, 'Hello'), (2, 'World');

WITH

WITH [ RECURSIVE ] { name [( columnName [,...] )] AS ( select ) [,...] }
{ select | insert | update | merge | delete | createTable }

Can be used to create a recursive or non-recursive query (common table expression). For recursive queries the first select has to be a UNION. One or more common table entries can be referred to by name. Column name declarations are now optional - the column names will be inferred from the named select queries. The final action in a WITH statement can be a select, insert, update, merge, delete or create table.

Example:

WITH RECURSIVE cte(n) AS (
        SELECT 1
    UNION ALL
        SELECT n + 1
        FROM cte
        WHERE n < 100
)
SELECT sum(n) FROM cte;

Example 2:
WITH cte1 AS (
        SELECT 1 AS FIRST_COLUMN
), cte2 AS (
        SELECT FIRST_COLUMN+1 AS FIRST_COLUMN FROM cte1
)
SELECT sum(FIRST_COLUMN) FROM cte2;

Commands (Data Definition)

ALTER INDEX RENAME

ALTER INDEX [ IF EXISTS ] indexName RENAME TO newIndexName

Renames an index. This command commits an open transaction in this connection.

Example:

ALTER INDEX IDXNAME RENAME TO IDX_TEST_NAME

ALTER SCHEMA RENAME

ALTER SCHEMA [ IF EXISTS ] schema RENAME TO newSchemaName

Renames a schema. This command commits an open transaction in this connection.

Example:

ALTER SCHEMA TEST RENAME TO PRODUCTION

ALTER SEQUENCE

ALTER SEQUENCE [ IF EXISTS ] sequenceName
[ RESTART WITH long ]
[ INCREMENT BY long ]
[ MINVALUE long | NOMINVALUE | NO MINVALUE ]
[ MAXVALUE long | NOMAXVALUE | NO MAXVALUE ]
[ CYCLE long | NOCYCLE | NO CYCLE ]
[ CACHE long | NOCACHE | NO CACHE ]

Changes the parameters of a sequence. This command does not commit the current transaction; however the new value is used by other transactions immediately, and rolling back this command has no effect.

Example:

ALTER SEQUENCE SEQ_ID RESTART WITH 1000

ALTER TABLE ADD

ALTER TABLE [ IF EXISTS ] tableName ADD [ COLUMN ]
{ [ IF NOT EXISTS ] columnName columnDefinition
| ( { columnName columnDefinition | constraint } [,...] ) }
[ { { BEFORE | AFTER } columnName } | FIRST ]

Adds a new column to a table. This command commits an open transaction in this connection.

Example:

ALTER TABLE TEST ADD CREATEDATE TIMESTAMP

ALTER TABLE ADD CONSTRAINT

ALTER TABLE [ IF EXISTS ] tableName ADD constraint [ CHECK | NOCHECK ]

Adds a constraint to a table. If NOCHECK is specified, existing rows are not checked for consistency (the default is to check consistency for existing rows). The required indexes are automatically created if they don't exist yet. It is not possible to disable checking for unique constraints. This command commits an open transaction in this connection.

Example:

ALTER TABLE TEST ADD CONSTRAINT NAME_UNIQUE UNIQUE(NAME)

ALTER TABLE RENAME CONSTRAINT

ALTER TABLE [ IF EXISTS ] tableName RENAME oldConstraintName
TO newConstraintName

Renames a constraint. This command commits an open transaction in this connection.

Example:

ALTER TABLE TEST RENAME CONSTRAINT FOO TO BAR

ALTER TABLE ALTER COLUMN

ALTER TABLE [ IF EXISTS ] tableName ALTER COLUMN columnName
{ { columnDefinition }
| { RENAME TO name }
| { RESTART WITH long }
| { SELECTIVITY int }
| { SET DEFAULT expression }
| { DROP DEFAULT }
| { SET ON UPDATE expression }
| { DROP ON UPDATE }
| { SET NOT NULL }
| { DROP NOT NULL } | { SET NULL }
| { SET DATA TYPE dataType }
| { SET { VISIBLE | INVISIBLE } } }

Changes the data type of a column, rename a column, change the identity value, or change the selectivity.

Changing the data type fails if the data can not be converted.

RESTART changes the next value of an auto increment column. The column must already be an auto increment column. For RESTART, the same transactional rules as for ALTER SEQUENCE apply.

SELECTIVITY sets the selectivity (1-100) for a column. Setting the selectivity to 0 means the default value. Selectivity is used by the cost based optimizer to calculate the estimated cost of an index. Selectivity 100 means values are unique, 10 means every distinct value appears 10 times on average.

SET DEFAULT changes the default value of a column.

DROP DEFAULT removes the default value of a column.

SET ON UPDATE changes the value that is set on update if value for this column is not specified in update statement.

DROP ON UPDATE removes the value that is set on update of a column.

SET NOT NULL sets a column to not allow NULL. Rows may not contains NULL in this column.

DROP NOT NULL and SET NULL set a column to allow NULL. The row may not be part of a primary key.

SET DATA TYPE changes the data type of a column.

SET INVISIBLE makes the column hidden, i.e. it will not appear in SELECT * results. SET VISIBLE has the reverse effect.

This command commits an open transaction in this connection.

Example:

ALTER TABLE TEST ALTER COLUMN NAME CLOB;
ALTER TABLE TEST ALTER COLUMN NAME RENAME TO TEXT;
ALTER TABLE TEST ALTER COLUMN ID RESTART WITH 10000;
ALTER TABLE TEST ALTER COLUMN NAME SELECTIVITY 100;
ALTER TABLE TEST ALTER COLUMN NAME SET DEFAULT '';
ALTER TABLE TEST ALTER COLUMN NAME SET NOT NULL;
ALTER TABLE TEST ALTER COLUMN NAME SET NULL;
ALTER TABLE TEST ALTER COLUMN NAME SET VISIBLE;
ALTER TABLE TEST ALTER COLUMN NAME SET INVISIBLE;

ALTER TABLE DROP COLUMN

ALTER TABLE [ IF EXISTS ] tableName DROP COLUMN [ IF EXISTS ]
columnName [,...] | ( columnName [,...] )

Removes column(s) from a table. This command commits an open transaction in this connection.

Example:

ALTER TABLE TEST DROP COLUMN NAME
ALTER TABLE TEST DROP COLUMN NAME1, NAME2
ALTER TABLE TEST DROP COLUMN (NAME1, NAME2)

ALTER TABLE DROP CONSTRAINT

ALTER TABLE [ IF EXISTS ] tableName DROP
{ CONSTRAINT [ IF EXISTS ] constraintName | PRIMARY KEY }

Removes a constraint or a primary key from a table. This command commits an open transaction in this connection.

Example:

ALTER TABLE TEST DROP CONSTRAINT UNIQUE_NAME

ALTER TABLE SET

ALTER TABLE [ IF EXISTS ] tableName SET REFERENTIAL_INTEGRITY
{ FALSE | TRUE } [ CHECK | NOCHECK ]

Disables or enables referential integrity checking for a table. This command can be used inside a transaction. Enabling referential integrity does not check existing data, except if CHECK is specified. Use SET REFERENTIAL_INTEGRITY to disable it for all tables; the global flag and the flag for each table are independent.

This command commits an open transaction in this connection.

Example:

ALTER TABLE TEST SET REFERENTIAL_INTEGRITY FALSE

ALTER TABLE RENAME

ALTER TABLE [ IF EXISTS ] tableName RENAME TO newName

Renames a table. This command commits an open transaction in this connection.

Example:

ALTER TABLE TEST RENAME TO MY_DATA

ALTER USER ADMIN

ALTER USER userName ADMIN { TRUE | FALSE }

Switches the admin flag of a user on or off.

Only unquoted or uppercase user names are allowed. Admin rights are required to execute this command. This command commits an open transaction in this connection.

Example:

ALTER USER TOM ADMIN TRUE

ALTER USER RENAME

ALTER USER userName RENAME TO newUserName

Renames a user. After renaming a user, the password becomes invalid and needs to be changed as well.

Only unquoted or uppercase user names are allowed. Admin rights are required to execute this command. This command commits an open transaction in this connection.

Example:

ALTER USER TOM RENAME TO THOMAS

ALTER USER SET PASSWORD

ALTER USER userName SET { PASSWORD string | SALT bytes HASH bytes }

Changes the password of a user. Only unquoted or uppercase user names are allowed. The password must be enclosed in single quotes. It is case sensitive and can contain spaces. The salt and hash values are hex strings.

Admin rights are required to execute this command. This command commits an open transaction in this connection.

Example:

ALTER USER SA SET PASSWORD 'rioyxlgt'

ALTER VIEW RECOMPILE

ALTER VIEW [ IF EXISTS ] viewName RECOMPILE

Recompiles a view after the underlying tables have been changed or created. This command is used for views created using CREATE FORCE VIEW. This command commits an open transaction in this connection.

Example:

ALTER VIEW ADDRESS_VIEW RECOMPILE

ALTER VIEW RENAME

ALTER VIEW [ IF EXISTS ] viewName RENAME TO newName

Renames a view. This command commits an open transaction in this connection.

Example:

ALTER VIEW TEST RENAME TO MY_VIEW

ANALYZE

ANALYZE [ TABLE tableName ] [ SAMPLE_SIZE rowCountInt ]

Updates the selectivity statistics of tables. If no table name is given, all tables are analyzed. The selectivity is used by the cost based optimizer to select the best index for a given query. If no sample size is set, up to 10000 rows per table are read. The value 0 means all rows are read. The selectivity can be set manually using ALTER TABLE ALTER COLUMN SELECTIVITY. Manual values are overwritten by this statement. The selectivity is available in the INFORMATION_SCHEMA.COLUMNS table.

This command commits an open transaction in this connection.

Example:

ANALYZE SAMPLE_SIZE 1000

COMMENT

COMMENT ON
{ { COLUMN [ schemaName. ] tableName.columnName }
| { { TABLE | VIEW | CONSTANT | CONSTRAINT | ALIAS | INDEX | ROLE
| SCHEMA | SEQUENCE | TRIGGER | USER | DOMAIN } [ schemaName. ] objectName } }
IS expression

Sets the comment of a database object. Use NULL to remove the comment.

Admin rights are required to execute this command. This command commits an open transaction in this connection.

Example:

COMMENT ON TABLE TEST IS 'Table used for testing'

CREATE AGGREGATE

CREATE AGGREGATE [ IF NOT EXISTS ] newAggregateName FOR className

Creates a new user-defined aggregate function. The method name must be the full qualified class name. The class must implement the interface org.h2.api.Aggregate or org.h2.api.AggregateFunction.

Admin rights are required to execute this command. This command commits an open transaction in this connection.

Example:

CREATE AGGREGATE SIMPLE_MEDIAN FOR "com.acme.db.Median"

CREATE ALIAS

CREATE ALIAS [ IF NOT EXISTS ] newFunctionAliasName [ DETERMINISTIC ]
[ NOBUFFER ] { FOR classAndMethodName | AS sourceCodeString }

Creates a new function alias. If this is a ResultSet returning function, by default the return value is cached in a local temporary file.

NOBUFFER - disables caching of ResultSet return value to temporary file.

DETERMINISTIC - Deterministic functions must always return the same value for the same parameters.

The method name must be the full qualified class and method name, and may optionally include the parameter classes as in java.lang.Integer.parseInt(java.lang.String, int). The class and the method must both be public, and the method must be static. The class must be available in the classpath of the database engine (when using the server mode, it must be in the classpath of the server).

When defining a function alias with source code, the Sun javac is compiler is used if the file tools.jar is in the classpath. If not, javac is run as a separate process. Only the source code is stored in the database; the class is compiled each time the database is re-opened. Source code is usually passed as dollar quoted text to avoid escaping problems. If import statements are used, then the tag @CODE must be added before the method.

If the method throws an SQLException, it is directly re-thrown to the calling application; all other exceptions are first converted to a SQLException.

If the first parameter of the Java function is a java.sql.Connection, then a connection to the database is provided. This connection must not be closed. If the class contains multiple methods with the given name but different parameter count, all methods are mapped.

Admin rights are required to execute this command. This command commits an open transaction in this connection.

If you have the Groovy jar in your classpath, it is also possible to write methods using Groovy.

Example:

CREATE ALIAS MY_SQRT FOR "java.lang.Math.sqrt";
CREATE ALIAS GET_SYSTEM_PROPERTY FOR "java.lang.System.getProperty";
CALL GET_SYSTEM_PROPERTY('java.class.path');
CALL GET_SYSTEM_PROPERTY('com.acme.test', 'true');
CREATE ALIAS REVERSE AS $$ String reverse(String s) { return new StringBuilder(s).reverse().toString(); } $$;
CALL REVERSE('Test');
CREATE ALIAS tr AS $$@groovy.transform.CompileStatic
    static String tr(String str, String sourceSet, String replacementSet){
        return str.tr(sourceSet, replacementSet);
    }
$$

CREATE CONSTANT

CREATE CONSTANT [ IF NOT EXISTS ] newConstantName VALUE expression

Creates a new constant. This command commits an open transaction in this connection.

Example:

CREATE CONSTANT ONE VALUE 1

CREATE DOMAIN

CREATE DOMAIN [ IF NOT EXISTS ] newDomainName AS dataType
[ DEFAULT expression ] [ [ NOT ] NULL ] [ SELECTIVITY selectivity ]
[ CHECK condition ]

Creates a new data type (domain). The check condition must evaluate to true or to NULL (to prevent NULL, use NOT NULL). In the condition, the term VALUE refers to the value being tested.

Domains are usable within the whole database. They can not be created in a specific schema.

This command commits an open transaction in this connection.

Example:

CREATE DOMAIN EMAIL AS VARCHAR(255) CHECK (POSITION('@', VALUE) > 1)

CREATE INDEX

CREATE
{ [ UNIQUE ] [ HASH | SPATIAL] INDEX [ [ IF NOT EXISTS ] newIndexName ]
| PRIMARY KEY [ HASH ] }
ON tableName ( indexColumn [,...] )

Creates a new index. This command commits an open transaction in this connection.

Hash indexes are meant for in-memory databases and memory tables (CREATE MEMORY TABLE) when PageStore engine is used. For other tables, or if the index contains multiple columns, the HASH keyword is ignored. Hash indexes can only test for equality, do not support range queries (similar to a hash table), use more memory, but can perform lookups faster. Non-unique keys are supported.

Spatial indexes are supported only on Geometry columns.

Example:

CREATE INDEX IDXNAME ON TEST(NAME)

CREATE LINKED TABLE

CREATE [ FORCE ] [ [ GLOBAL | LOCAL ] TEMPORARY ]
LINKED TABLE [ IF NOT EXISTS ]
name ( driverString, urlString, userString, passwordString,
[ originalSchemaString, ] originalTableString ) [ EMIT UPDATES | READONLY ]

Creates a table link to an external table. The driver name may be empty if the driver is already loaded. If the schema name is not set, only one table with that name may exist in the target database.

FORCE - Create the LINKED TABLE even if the remote database/table does not exist.

EMIT UPDATES - Usually, for update statements, the old rows are deleted first and then the new rows are inserted. It is possible to emit update statements (except on rollback), however in this case multi-row unique key updates may not always work. Linked tables to the same database share one connection.

READONLY - is set, the remote table may not be updated. This is enforced by H2.

If the connection to the source database is lost, the connection is re-opened (this is a workaround for MySQL that disconnects after 8 hours of inactivity by default).

If a query is used instead of the original table name, the table is read only. Queries must be enclosed in parenthesis: (SELECT * FROM ORDERS).

To use JNDI to get the connection, the driver class must be a javax.naming.Context (for example javax.naming.InitialContext), and the URL must be the resource name (for example java:comp/env/jdbc/Test).

Admin rights are required to execute this command. This command commits an open transaction in this connection.

Example:

CREATE LINKED TABLE LINK('org.h2.Driver', 'jdbc:h2:test2',
    'sa', 'sa', 'TEST');
CREATE LINKED TABLE LINK('', 'jdbc:h2:test2', 'sa', 'sa',
    '(SELECT * FROM TEST WHERE ID>0)');
CREATE LINKED TABLE LINK('javax.naming.InitialContext',
    'java:comp/env/jdbc/Test', NULL, NULL,
    '(SELECT * FROM TEST WHERE ID>0)');

CREATE ROLE

CREATE ROLE [ IF NOT EXISTS ] newRoleName

Creates a new role. This command commits an open transaction in this connection.

Example:

CREATE ROLE READONLY

CREATE SCHEMA

CREATE SCHEMA [ IF NOT EXISTS ] name
[ AUTHORIZATION ownerUserName ]
[ WITH tableEngineParamName [,...] ]

Creates a new schema. If no owner is specified, the current user is used. The user that executes the command must have admin rights, as well as the owner. Specifying the owner currently has no effect. Optional table engine parameters are used when CREATE TABLE command is run on this schema without having its engine params set.

This command commits an open transaction in this connection.

Example:

CREATE SCHEMA TEST_SCHEMA AUTHORIZATION SA

CREATE SEQUENCE

CREATE SEQUENCE [ IF NOT EXISTS ] newSequenceName
sequenceOptions

Creates a new sequence. The data type of a sequence is BIGINT. Used values are never re-used, even when the transaction is rolled back.

The cache is the number of pre-allocated numbers. If the system crashes without closing the database, at most this many numbers are lost. The default cache size is 32. To disable caching, use the cache size 1 or lower.

This command commits an open transaction in this connection.

Example:

CREATE SEQUENCE SEQ_ID

CREATE TABLE

CREATE [ CACHED | MEMORY ] [ TEMP | [ GLOBAL | LOCAL ] TEMPORARY ]
TABLE [ IF NOT EXISTS ] name
[ ( { columnName [columnDefinition] | constraint } [,...] ) ]
[ ENGINE tableEngineName ]
[ WITH tableEngineParamName [,...] ]
[ NOT PERSISTENT ] [ TRANSACTIONAL ]
[ AS select [ WITH [ NO ] DATA ] ]

Creates a new table.

Cached tables (the default for regular tables) are persistent, and the number of rows is not limited by the main memory. Memory tables (the default for temporary tables) are persistent, but the index data is kept in main memory, that means memory tables should not get too large.

Temporary tables are deleted when closing or opening a database. Temporary tables can be global (accessible by all connections) or local (only accessible by the current connection). The default for temporary tables is global. Indexes of temporary tables are kept fully in main memory, unless the temporary table is created using CREATE CACHED TABLE.

The ENGINE option is only required when custom table implementations are used. The table engine class must implement the interface org.h2.api.TableEngine. Any table engine parameters are passed down in the tableEngineParams field of the CreateTableData object.

Either ENGINE, or WITH (table engine params), or both may be specified. If ENGINE is not specified in CREATE TABLE, then the engine specified by DEFAULT_TABLE_ENGINE option of database params is used.

Tables with the NOT PERSISTENT modifier are kept fully in memory, and all rows are lost when the database is closed.

The column definitions are optional if a query is specified. In that case the column list of the query is used. If the query is specified its results are inserted into created table unless WITH NO DATA is specified.

This command commits an open transaction, except when using TRANSACTIONAL (only supported for temporary tables).

Example:

CREATE TABLE TEST(ID INT PRIMARY KEY, NAME VARCHAR(255))

CREATE TRIGGER

CREATE TRIGGER [ IF NOT EXISTS ] newTriggerName
{ BEFORE | AFTER | INSTEAD OF }
{ INSERT | UPDATE | DELETE | SELECT | ROLLBACK }
[,...] ON tableName [ FOR EACH ROW ]
[ QUEUE int ] [ NOWAIT ]
{ CALL triggeredClassName | AS sourceCodeString }

Creates a new trigger. The trigger class must be public and implement org.h2.api.Trigger. Inner classes are not supported. The class must be available in the classpath of the database engine (when using the server mode, it must be in the classpath of the server).

The sourceCodeString must define a single method with no parameters that returns org.h2.api.Trigger. See CREATE ALIAS for requirements regarding the compilation. Alternatively, javax.script.ScriptEngineManager can be used to create an instance of org.h2.api.Trigger. Currently javascript (included in every JRE) and ruby (with JRuby) are supported. In that case the source must begin respectively with //javascript or #ruby.

BEFORE triggers are called after data conversion is made, default values are set, null and length constraint checks have been made; but before other constraints have been checked. If there are multiple triggers, the order in which they are called is undefined.

ROLLBACK can be specified in combination with INSERT, UPDATE, and DELETE. Only row based AFTER trigger can be called on ROLLBACK. Exceptions that occur within such triggers are ignored. As the operations that occur within a trigger are part of the transaction, ROLLBACK triggers are only required if an operation communicates outside of the database.

INSTEAD OF triggers are implicitly row based and behave like BEFORE triggers. Only the first such trigger is called. Such triggers on views are supported. They can be used to make views updatable.

A BEFORE SELECT trigger is fired just before the database engine tries to read from the table. The trigger can be used to update a table on demand. The trigger is called with both 'old' and 'new' set to null.

The MERGE statement will call both INSERT and UPDATE triggers. Not supported are SELECT triggers with the option FOR EACH ROW, and AFTER SELECT triggers.

Committing or rolling back a transaction within a trigger is not allowed, except for SELECT triggers.

By default a trigger is called once for each statement, without the old and new rows. FOR EACH ROW triggers are called once for each inserted, updated, or deleted row.

QUEUE is implemented for syntax compatibility with HSQL and has no effect.

The trigger need to be created in the same schema as the table. The schema name does not need to be specified when creating the trigger.

This command commits an open transaction in this connection.

Example:

CREATE TRIGGER TRIG_INS BEFORE INSERT ON TEST FOR EACH ROW CALL "MyTrigger";
CREATE TRIGGER TRIG_SRC BEFORE INSERT ON TEST AS $$org.h2.api.Trigger create() { return new MyTrigger("constructorParam"); } $$;
CREATE TRIGGER TRIG_JS BEFORE INSERT ON TEST AS $$//javascript\nreturn new Packages.MyTrigger("constructorParam"); $$;
CREATE TRIGGER TRIG_RUBY BEFORE INSERT ON TEST AS $$#ruby\nJava::MyPackage::MyTrigger.new("constructorParam") $$;

CREATE USER

CREATE USER [ IF NOT EXISTS ] newUserName
{ PASSWORD string | SALT bytes HASH bytes } [ ADMIN ]

Creates a new user. For compatibility, only unquoted or uppercase user names are allowed. The password must be in single quotes. It is case sensitive and can contain spaces. The salt and hash values are hex strings.

Admin rights are required to execute this command. This command commits an open transaction in this connection.

Example:

CREATE USER GUEST PASSWORD 'abc'

CREATE VIEW

CREATE [ OR REPLACE ] [ FORCE ] VIEW [ IF NOT EXISTS ] newViewName
[ ( columnName [,...] ) ] AS select

Creates a new view. If the force option is used, then the view is created even if the underlying table(s) don't exist.

If the OR REPLACE clause is used an existing view will be replaced, and any dependent views will not need to be recreated. If dependent views will become invalid as a result of the change an error will be generated, but this error can be ignored if the FORCE clause is also used.

Views are not updatable except when using 'instead of' triggers.

Admin rights are required to execute this command. This command commits an open transaction in this connection.

Example:

CREATE VIEW TEST_VIEW AS SELECT * FROM TEST WHERE ID < 100

DROP AGGREGATE

DROP AGGREGATE [ IF EXISTS ] aggregateName

Drops an existing user-defined aggregate function.

Admin rights are required to execute this command. This command commits an open transaction in this connection.

Example:

DROP AGGREGATE SIMPLE_MEDIAN

DROP ALIAS

DROP ALIAS [ IF EXISTS ] existingFunctionAliasName

Drops an existing function alias.

Admin rights are required to execute this command. This command commits an open transaction in this connection.

Example:

DROP ALIAS MY_SQRT

DROP ALL OBJECTS

DROP ALL OBJECTS [ DELETE FILES ]

Drops all existing views, tables, sequences, schemas, function aliases, roles, user-defined aggregate functions, domains, and users (except the current user). If DELETE FILES is specified, the database files will be removed when the last user disconnects from the database. Warning: this command can not be rolled back.

Admin rights are required to execute this command.

Example:

DROP ALL OBJECTS

DROP CONSTANT

DROP CONSTANT [ IF EXISTS ] constantName

Drops a constant. This command commits an open transaction in this connection.

Example:

DROP CONSTANT ONE

DROP DOMAIN

DROP DOMAIN [ IF EXISTS ] domainName [ RESTRICT | CASCADE ]

Drops a data type (domain). The command will fail if it is referenced by a column (the default). Column descriptors are replaced with original definition of specified domain if the CASCADE clause is used. This command commits an open transaction in this connection.

Example:

DROP DOMAIN EMAIL

DROP INDEX

DROP INDEX [ IF EXISTS ] indexName

Drops an index. This command commits an open transaction in this connection.

Example:

DROP INDEX IF EXISTS IDXNAME

DROP ROLE

DROP ROLE [ IF EXISTS ] roleName

Drops a role. This command commits an open transaction in this connection.

Example:

DROP ROLE READONLY

DROP SCHEMA

DROP SCHEMA [ IF EXISTS ] schemaName [ RESTRICT | CASCADE ]

Drops a schema. The command will fail if objects in this schema exist and the RESTRICT clause is used (the default). All objects in this schema are dropped as well if the CASCADE clause is used. This command commits an open transaction in this connection.

Example:

DROP SCHEMA TEST_SCHEMA

DROP SEQUENCE

DROP SEQUENCE [ IF EXISTS ] sequenceName

Drops a sequence. This command commits an open transaction in this connection.

Example:

DROP SEQUENCE SEQ_ID

DROP TABLE

DROP TABLE [ IF EXISTS ] tableName [,...] [ RESTRICT | CASCADE ]

Drops an existing table, or a list of tables. The command will fail if dependent objects exist and the RESTRICT clause is used (the default). All dependent views and constraints are dropped as well if the CASCADE clause is used. This command commits an open transaction in this connection.

Example:

DROP TABLE TEST

DROP TRIGGER

DROP TRIGGER [ IF EXISTS ] triggerName

Drops an existing trigger. This command commits an open transaction in this connection.

Example:

DROP TRIGGER TRIG_INS

DROP USER

DROP USER [ IF EXISTS ] userName

Drops a user. The current user cannot be dropped. For compatibility, only unquoted or uppercase user names are allowed.

Admin rights are required to execute this command. This command commits an open transaction in this connection.

Example:

DROP USER TOM

DROP VIEW

DROP VIEW [ IF EXISTS ] viewName [ RESTRICT | CASCADE ]

Drops an existing view. All dependent views are dropped as well if the CASCADE clause is used (the default). The command will fail if dependent views exist and the RESTRICT clause is used. This command commits an open transaction in this connection.

Example:

DROP VIEW TEST_VIEW

TRUNCATE TABLE

TRUNCATE TABLE tableName [ [ CONTINUE | RESTART ] IDENTITY ]

Removes all rows from a table. Unlike DELETE FROM without where clause, this command can not be rolled back. This command is faster than DELETE without where clause. Only regular data tables without foreign key constraints can be truncated (except if referential integrity is disabled for this database or for this table). Linked tables can't be truncated. If RESTART IDENTITY is specified next values for auto-incremented columns are restarted.

This command commits an open transaction in this connection.

Example:

TRUNCATE TABLE TEST

Commands (Other)

CHECKPOINT

CHECKPOINT

Flushes the data to disk.

Admin rights are required to execute this command.

Example:

CHECKPOINT

CHECKPOINT SYNC

CHECKPOINT SYNC

Flushes the data to disk and forces all system buffers be written to the underlying device.

Admin rights are required to execute this command.

Example:

CHECKPOINT SYNC

COMMIT

COMMIT [ WORK ]

Commits a transaction.

Example:

COMMIT

COMMIT TRANSACTION

COMMIT TRANSACTION transactionName

Sets the resolution of an in-doubt transaction to 'commit'.

Admin rights are required to execute this command. This command is part of the 2-phase-commit protocol.

Example:

COMMIT TRANSACTION XID_TEST

GRANT RIGHT

GRANT { SELECT | INSERT | UPDATE | DELETE | ALL } [,...] ON
{ { SCHEMA schemaName } | { tableName [,...] } }
TO { PUBLIC | userName | roleName }

Grants rights for a table to a user or role.

Admin rights are required to execute this command. This command commits an open transaction in this connection.

Example:

GRANT SELECT ON TEST TO READONLY

GRANT ALTER ANY SCHEMA

GRANT ALTER ANY SCHEMA TO userName

Grant schema altering rights to a user.

Admin rights are required to execute this command. This command commits an open transaction in this connection.

Example:

GRANT ALTER ANY SCHEMA TO Bob

GRANT ROLE

GRANT roleName TO { PUBLIC | userName | roleName }

Grants a role to a user or role.

Admin rights are required to execute this command. This command commits an open transaction in this connection.

Example:

GRANT READONLY TO PUBLIC

HELP

HELP [ anything [...] ]

Displays the help pages of SQL commands or keywords.

Example:

HELP SELECT

PREPARE COMMIT

PREPARE COMMIT newTransactionName

Prepares committing a transaction. This command is part of the 2-phase-commit protocol.

Example:

PREPARE COMMIT XID_TEST

REVOKE RIGHT

REVOKE { SELECT | INSERT | UPDATE | DELETE | ALL } [,...] ON
{ { SCHEMA schemaName } | { tableName [,...] } }
FROM { PUBLIC | userName | roleName }

Removes rights for a table from a user or role.

Admin rights are required to execute this command. This command commits an open transaction in this connection.

Example:

REVOKE SELECT ON TEST FROM READONLY

REVOKE ROLE

REVOKE roleName FROM { PUBLIC | userName | roleName }

Removes a role from a user or role.

Admin rights are required to execute this command. This command commits an open transaction in this connection.

Example:

REVOKE READONLY FROM TOM

ROLLBACK

ROLLBACK [ TO SAVEPOINT savepointName ]

Rolls back a transaction. If a savepoint name is used, the transaction is only rolled back to the specified savepoint.

Example:

ROLLBACK

ROLLBACK TRANSACTION

ROLLBACK TRANSACTION transactionName

Sets the resolution of an in-doubt transaction to 'rollback'.

Admin rights are required to execute this command. This command is part of the 2-phase-commit protocol.

Example:

ROLLBACK TRANSACTION XID_TEST

SAVEPOINT

SAVEPOINT savepointName

Create a new savepoint. See also ROLLBACK. Savepoints are only valid until the transaction is committed or rolled back.

Example:

SAVEPOINT HALF_DONE

SET @

SET @variableName [ = ] expression

Updates a user-defined variable. Variables are not persisted and session scoped, that means only visible from within the session in which they are defined. This command does not commit a transaction, and rollback does not affect it.

Example:

SET @TOTAL=0

SET ALLOW_LITERALS

SET ALLOW_LITERALS { NONE | ALL | NUMBERS }

This setting can help solve the SQL injection problem. By default, text and number literals are allowed in SQL statements. However, this enables SQL injection if the application dynamically builds SQL statements. SQL injection is not possible if user data is set using parameters ('?').

NONE means literals of any kind are not allowed, only parameters and constants are allowed. NUMBERS mean only numerical and boolean literals are allowed. ALL means all literals are allowed (default).

See also CREATE CONSTANT.

Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting is persistent. This setting can be appended to the database URL: jdbc:h2:test;ALLOW_LITERALS=NONE

Example:

SET ALLOW_LITERALS NONE

SET AUTOCOMMIT

SET AUTOCOMMIT { TRUE | ON | FALSE | OFF }

Switches auto commit on or off. This setting can be appended to the database URL: jdbc:h2:test;AUTOCOMMIT=OFF - however this will not work as expected when using a connection pool (the connection pool manager will re-enable autocommit when returning the connection to the pool, so autocommit will only be disabled the first time the connection is used.

Example:

SET AUTOCOMMIT OFF

SET CACHE_SIZE

SET CACHE_SIZE int

Sets the size of the cache in KB (each KB being 1024 bytes) for the current database. The default is 65536 per available GB of RAM, i.e. 64 MB per GB. The value is rounded to the next higher power of two. Depending on the virtual machine, the actual memory required may be higher.

This setting is persistent and affects all connections as there is only one cache per database. Using a very small value (specially 0) will reduce performance a lot. This setting only affects the database engine (the server in a client/server environment; in embedded mode, the database engine is in the same process as the application). It has no effect for in-memory databases.

Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting is persistent. This setting can be appended to the database URL: jdbc:h2:test;CACHE_SIZE=8192

Example:

SET CACHE_SIZE 8192

SET CLUSTER

SET CLUSTER serverListString

This command should not be used directly by an application, the statement is executed automatically by the system. The behavior may change in future releases. Sets the cluster server list. An empty string switches off the cluster mode. Switching on the cluster mode requires admin rights, but any user can switch it off (this is automatically done when the client detects the other server is not responding).

This command is effective immediately, but does not commit an open transaction.

Example:

SET CLUSTER ''

SET BINARY_COLLATION

SET BINARY_COLLATION { UNSIGNED | SIGNED }

Sets the collation used for comparing BINARY columns, the default is SIGNED for version 1.3 and older, and UNSIGNED for version 1.4 and newer. This command can only be executed if there are no tables defined.

Admin rights are required to execute this command. This command commits an open transaction in this connection. This setting is persistent.

Example:

SET BINARY_COLLATION SIGNED

SET UUID_COLLATION

SET UUID_COLLATION { UNSIGNED | SIGNED }

Sets the collation used for comparing UUID columns, the default is SIGNED. This command can only be executed if there are no tables defined.

SIGNED means signed comparison between first 64 bits of compared values treated as long values and if they are equal a signed comparison of the last 64 bits of compared values treated as long values. See also Java UUID.compareTo(). UNSIGNED means RFC 4122 compatible unsigned comparison.

Admin rights are required to execute this command. This command commits an open transaction in this connection. This setting is persistent.

Example:

SET UUID_COLLATION UNSIGNED

SET BUILTIN_ALIAS_OVERRIDE

SET BUILTIN_ALIAS_OVERRIDE { TRUE | FALSE }

Allows the overriding of the builtin system date/time functions for unit testing purposes.

Admin rights are required to execute this command. This command commits an open transaction in this connection.

Example:

SET BUILTIN_ALIAS_OVERRIDE TRUE

SET COLLATION

SET [ DATABASE ] COLLATION
{ OFF | collationName
[ STRENGTH { PRIMARY | SECONDARY | TERTIARY | IDENTICAL } ] }

Sets the collation used for comparing strings. This command can only be executed if there are no tables defined. See java.text.Collator for details about the supported collations and the STRENGTH (PRIMARY is usually case- and umlaut-insensitive; SECONDARY is case-insensitive but umlaut-sensitive; TERTIARY is both case- and umlaut-sensitive; IDENTICAL is sensitive to all differences and only affects ordering).

The ICU4J collator is used if it is in the classpath. It is also used if the collation name starts with ICU4J_ (in that case, the ICU4J must be in the classpath, otherwise an exception is thrown). The default collator is used if the collation name starts with DEFAULT_ (even if ICU4J is in the classpath). The charset collator is used if the collation name starts with CHARSET_ (e.g. CHARSET_CP500). This collator sorts strings according to the binary representation in the given charset.

Admin rights are required to execute this command. This command commits an open transaction in this connection. This setting is persistent. This setting can be appended to the database URL: jdbc:h2:test;COLLATION='ENGLISH'

Example:

SET COLLATION ENGLISH
SET COLLATION CHARSET_CP500

SET COMPRESS_LOB

SET COMPRESS_LOB { NO | LZF | DEFLATE }

This feature is only available for the PageStore storage engine. For the MVStore engine (the default for H2 version 1.4.x), append ;COMPRESS=TRUE to the database URL instead.

Sets the compression algorithm for BLOB and CLOB data. Compression is usually slower, but needs less disk space. LZF is faster but uses more space.

Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting is persistent.

Example:

SET COMPRESS_LOB LZF

SET DATABASE_EVENT_LISTENER

SET DATABASE_EVENT_LISTENER classNameString

Sets the event listener class. An empty string ('') means no listener should be used. This setting is not persistent.

Admin rights are required to execute this command, except if it is set when opening the database (in this case it is reset just after opening the database). This setting can be appended to the database URL: jdbc:h2:test;DATABASE_EVENT_LISTENER='sample.MyListener'

Example:

SET DATABASE_EVENT_LISTENER 'sample.MyListener'

SET DB_CLOSE_DELAY

SET DB_CLOSE_DELAY int

Sets the delay for closing a database if all connections are closed. The value -1 means the database is never closed until the close delay is set to some other value or SHUTDOWN is called. The value 0 means no delay (default; the database is closed if the last connection to it is closed). Values 1 and larger mean the number of seconds the database is left open after closing the last connection.

If the application exits normally or System.exit is called, the database is closed immediately, even if a delay is set.

Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting is persistent. This setting can be appended to the database URL: jdbc:h2:test;DB_CLOSE_DELAY=-1

Example:

SET DB_CLOSE_DELAY -1

SET DEFAULT_LOCK_TIMEOUT

SET DEFAULT LOCK_TIMEOUT int

Sets the default lock timeout (in milliseconds) in this database that is used for the new sessions. The default value for this setting is 1000 (one second).

Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting is persistent.

Example:

SET DEFAULT_LOCK_TIMEOUT 5000

SET DEFAULT_TABLE_TYPE

SET DEFAULT_TABLE_TYPE { MEMORY | CACHED }

Sets the default table storage type that is used when creating new tables. Memory tables are kept fully in the main memory (including indexes), however the data is still stored in the database file. The size of memory tables is limited by the memory. The default is CACHED.

Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting is persistent. It has no effect for in-memory databases.

Example:

SET DEFAULT_TABLE_TYPE MEMORY

SET EXCLUSIVE

SET EXCLUSIVE { 0 | 1 | 2 }

Switched the database to exclusive mode (1, 2) and back to normal mode (0).

In exclusive mode, new connections are rejected, and operations by other connections are paused until the exclusive mode is disabled. When using the value 1, existing connections stay open. When using the value 2, all existing connections are closed (and current transactions are rolled back) except the connection that executes SET EXCLUSIVE. Only the connection that set the exclusive mode can disable it. When the connection is closed, it is automatically disabled.

Admin rights are required to execute this command. This command commits an open transaction in this connection.

Example:

SET EXCLUSIVE 1

SET IGNORECASE

SET IGNORECASE { TRUE | FALSE }

If IGNORECASE is enabled, text columns in newly created tables will be case-insensitive. Already existing tables are not affected. The effect of case-insensitive columns is similar to using a collation with strength PRIMARY. Case-insensitive columns are compared faster than when using a collation. String literals and parameters are however still considered case sensitive even if this option is set.

Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting is persistent. This setting can be appended to the database URL: jdbc:h2:test;IGNORECASE=TRUE

Example:

SET IGNORECASE TRUE

SET JAVA_OBJECT_SERIALIZER

SET JAVA_OBJECT_SERIALIZER
{ null | className }

Sets the object used to serialize and deserialize java objects being stored in column of type OTHER. The serializer class must be public and implement org.h2.api.JavaObjectSerializer. Inner classes are not supported. The class must be available in the classpath of the database engine (when using the server mode, it must be both in the classpath of the server and the client). This command can only be executed if there are no tables defined.

Admin rights are required to execute this command. This command commits an open transaction in this connection. This setting is persistent. This setting can be appended to the database URL: jdbc:h2:test;JAVA_OBJECT_SERIALIZER='com.acme.SerializerClassName'

Example:

SET JAVA_OBJECT_SERIALIZER 'com.acme.SerializerClassName'

SET LAZY_QUERY_EXECUTION

SET LAZY_QUERY_EXECUTION int

Sets the lazy query execution mode. The values 0, 1 are supported.

If true, then large results are retrieved in chunks.

Note that not all queries support this feature, queries which do not are processed normally.

Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting is not persistent. This setting can be appended to the database URL: jdbc:h2:test;LAZY_QUERY_EXECUTION=1

Example:

SET LAZY_QUERY_EXECUTION 1

SET LOG

SET LOG int

Sets the transaction log mode. The values 0, 1, and 2 are supported, the default is 2. This setting affects all connections.

LOG 0 means the transaction log is disabled completely. It is the fastest mode, but also the most dangerous: if the process is killed while the database is open in this mode, the data might be lost. It must only be used if this is not a problem, for example when initially loading a database, or when running tests.

LOG 1 means the transaction log is enabled, but FileDescriptor.sync is disabled. This setting is about half as fast as with LOG 0. This setting is useful if no protection against power failure is required, but the data must be protected against killing the process.

LOG 2 (the default) means the transaction log is enabled, and FileDescriptor.sync is called for each checkpoint. This setting is about half as fast as LOG 1. Depending on the file system, this will also protect against power failure in the majority if cases.

Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting is not persistent. This setting can be appended to the database URL: jdbc:h2:test;LOG=0

Example:

SET LOG 1

SET LOCK_MODE

SET LOCK_MODE int

Sets the lock mode. The values 0, 1, 2, and 3 are supported. The default is 3 (READ_COMMITTED). This setting affects all connections.

The value 0 means no locking (should only be used for testing; also known as READ_UNCOMMITTED). Please note that using SET LOCK_MODE 0 while at the same time using multiple connections may result in inconsistent transactions.

The value 1 means table level locking (also known as SERIALIZABLE).

The value 2 means table level locking with garbage collection (if the application does not close all connections).

The value 3 means table level locking, but read locks are released immediately (default; also known as READ_COMMITTED).

Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting is persistent. This setting can be appended to the database URL: jdbc:h2:test;LOCK_MODE=3

Example:

SET LOCK_MODE 1

SET LOCK_TIMEOUT

SET LOCK_TIMEOUT int

Sets the lock timeout (in milliseconds) for the current session. The default value for this setting is 1000 (one second).

This command does not commit a transaction, and rollback does not affect it. This setting can be appended to the database URL: jdbc:h2:test;LOCK_TIMEOUT=10000

Example:

SET LOCK_TIMEOUT 1000

SET MAX_LENGTH_INPLACE_LOB

SET MAX_LENGTH_INPLACE_LOB int

Sets the maximum size of an in-place LOB object.

This is the maximum length of an LOB that is stored with the record itself, and the default value is 128.

Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting is persistent.

Example:

SET MAX_LENGTH_INPLACE_LOB 128

SET MAX_LOG_SIZE

SET MAX_LOG_SIZE int

Sets the maximum size of the transaction log, in megabytes. If the log is larger, and if there is no open transaction, the transaction log is truncated. If there is an open transaction, the transaction log will continue to grow however. The default max size is 16 MB. This setting has no effect for in-memory databases.

Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting is persistent.

Example:

SET MAX_LOG_SIZE 2

SET MAX_MEMORY_ROWS

SET MAX_MEMORY_ROWS int

The maximum number of rows in a result set that are kept in-memory. If more rows are read, then the rows are buffered to disk. The default is 40000 per GB of available RAM.

Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting is persistent. It has no effect for in-memory databases.

Example:

SET MAX_MEMORY_ROWS 1000

SET MAX_MEMORY_UNDO

SET MAX_MEMORY_UNDO int

The maximum number of undo records per a session that are kept in-memory. If a transaction is larger, the records are buffered to disk. The default value is 50000. Changes to tables without a primary key can not be buffered to disk. This setting is not supported when using multi-version concurrency.

Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting is persistent. It has no effect for in-memory databases.

Example:

SET MAX_MEMORY_UNDO 1000

SET MAX_OPERATION_MEMORY

SET MAX_OPERATION_MEMORY int

Sets the maximum memory used for large operations (delete and insert), in bytes. Operations that use more memory are buffered to disk, slowing down the operation. The default max size is 100000. 0 means no limit.

This setting is not persistent. Admin rights are required to execute this command, as it affects all connections. It has no effect for in-memory databases. This setting can be appended to the database URL: jdbc:h2:test;MAX_OPERATION_MEMORY=10000

Example:

SET MAX_OPERATION_MEMORY 0

SET MODE

SET MODE { REGULAR | DB2 | DERBY | HSQLDB | MSSQLSERVER | MYSQL | ORACLE | POSTGRESQL }

Changes to another database compatibility mode. For details, see Compatibility Modes in the feature section.

This setting is not persistent. Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting can be appended to the database URL: jdbc:h2:test;MODE=MYSQL

Example:

SET MODE HSQLDB

SET MULTI_THREADED

SET MULTI_THREADED { 0 | 1 }

Enabled (1) or disabled (0) multi-threading inside the database engine. MULTI_THREADED is enabled by default with default MVStore storage engine. MULTI_THREADED is disabled by default when using PageStore storage engine, enabling this with PageStore is experimental only.

This is a global setting, which means it is not possible to open multiple databases with different modes at the same time in the same virtual machine. This setting is not persistent, however the value is kept until the virtual machine exits or it is changed.

Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting can be appended to the database URL: jdbc:h2:test;MULTI_THREADED=1

Example:

SET MULTI_THREADED 1

SET OPTIMIZE_REUSE_RESULTS

SET OPTIMIZE_REUSE_RESULTS { 0 | 1 }

Enabled (1) or disabled (0) the result reuse optimization. If enabled, subqueries and views used as subqueries are only re-run if the data in one of the tables was changed. This option is enabled by default.

Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting can be appended to the database URL: jdbc:h2:test;OPTIMIZE_REUSE_RESULTS=0

Example:

SET OPTIMIZE_REUSE_RESULTS 0

SET PASSWORD

SET PASSWORD string

Changes the password of the current user. The password must be in single quotes. It is case sensitive and can contain spaces.

This command commits an open transaction in this connection.

Example:

SET PASSWORD 'abcstzri!.5'

SET QUERY_STATISTICS

SET QUERY_STATISTICS { TRUE | FALSE }

Disabled or enables query statistics gathering for the whole database. The statistics are reflected in the INFORMATION_SCHEMA.QUERY_STATISTICS meta-table.

This setting is not persistent. This command commits an open transaction in this connection. Admin rights are required to execute this command, as it affects all connections.

Example:

SET QUERY_STATISTICS FALSE

SET QUERY_STATISTICS_MAX_ENTRIES

SET QUERY_STATISTICS int

Set the maximum number of entries in query statistics meta-table. Default value is 100.

This setting is not persistent. This command commits an open transaction in this connection. Admin rights are required to execute this command, as it affects all connections.

Example:

SET QUERY_STATISTICS_MAX_ENTRIES 500

SET QUERY_TIMEOUT

SET QUERY_TIMEOUT int

Set the query timeout of the current session to the given value. The timeout is in milliseconds. All kinds of statements will throw an exception if they take longer than the given value. The default timeout is 0, meaning no timeout.

This command does not commit a transaction, and rollback does not affect it.

Example:

SET QUERY_TIMEOUT 10000

SET REFERENTIAL_INTEGRITY

SET REFERENTIAL_INTEGRITY { TRUE | FALSE }

Disabled or enables referential integrity checking for the whole database. Enabling it does not check existing data. Use ALTER TABLE SET to disable it only for one table.

This setting is not persistent. This command commits an open transaction in this connection. Admin rights are required to execute this command, as it affects all connections.

Example:

SET REFERENTIAL_INTEGRITY FALSE

SET RETENTION_TIME

SET RETENTION_TIME int

This property is only used when using the MVStore storage engine. How long to retain old, persisted data, in milliseconds. The default is 45000 (45 seconds), 0 means overwrite data as early as possible. It is assumed that a file system and hard disk will flush all write buffers within this time. Using a lower value might be dangerous, unless the file system and hard disk flush the buffers earlier. To manually flush the buffers, use CHECKPOINT SYNC, however please note that according to various tests this does not always work as expected depending on the operating system and hardware.

Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting is persistent. This setting can be appended to the database URL: jdbc:h2:test;RETENTION_TIME=0

Example:

SET RETENTION_TIME 0

SET SALT HASH

SET SALT bytes HASH bytes

Sets the password salt and hash for the current user. The password must be in single quotes. It is case sensitive and can contain spaces.

This command commits an open transaction in this connection.

Example:

SET SALT '00' HASH '1122'

SET SCHEMA

SET SCHEMA schemaName

Changes the default schema of the current connection. The default schema is used in statements where no schema is set explicitly. The default schema for new connections is PUBLIC.

This command does not commit a transaction, and rollback does not affect it. This setting can be appended to the database URL: jdbc:h2:test;SCHEMA=ABC

Example:

SET SCHEMA INFORMATION_SCHEMA

SET SCHEMA_SEARCH_PATH

SET SCHEMA_SEARCH_PATH schemaName [,...]

Changes the schema search path of the current connection. The default schema is used in statements where no schema is set explicitly. The default schema for new connections is PUBLIC.

This command does not commit a transaction, and rollback does not affect it. This setting can be appended to the database URL: jdbc:h2:test;SCHEMA_SEARCH_PATH=ABC,DEF

Example:

SET SCHEMA_SEARCH_PATH INFORMATION_SCHEMA, PUBLIC

SET THROTTLE

SET THROTTLE int

Sets the throttle for the current connection. The value is the number of milliseconds delay after each 50 ms. The default value is 0 (throttling disabled).

This command does not commit a transaction, and rollback does not affect it. This setting can be appended to the database URL: jdbc:h2:test;THROTTLE=50

Example:

SET THROTTLE 200

SET TRACE_LEVEL

SET { TRACE_LEVEL_FILE | TRACE_LEVEL_SYSTEM_OUT } int

Sets the trace level for file the file or system out stream. Levels are: 0=off, 1=error, 2=info, 3=debug. The default level is 1 for file and 0 for system out. To use SLF4J, append ;TRACE_LEVEL_FILE=4 to the database URL when opening the database.

This setting is not persistent. Admin rights are required to execute this command, as it affects all connections. This command does not commit a transaction, and rollback does not affect it. This setting can be appended to the database URL: jdbc:h2:test;TRACE_LEVEL_SYSTEM_OUT=3

Example:

SET TRACE_LEVEL_SYSTEM_OUT 3

SET TRACE_MAX_FILE_SIZE

SET TRACE_MAX_FILE_SIZE int

Sets the maximum trace file size. If the file exceeds the limit, the file is renamed to .old and a new file is created. If another .old file exists, it is deleted. The default max size is 16 MB.

This setting is persistent. Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting can be appended to the database URL: jdbc:h2:test;TRACE_MAX_FILE_SIZE=3

Example:

SET TRACE_MAX_FILE_SIZE 10

SET UNDO_LOG

SET UNDO_LOG int

Enables (1) or disables (0) the per session undo log. The undo log is enabled by default. When disabled, transactions can not be rolled back. This setting should only be used for bulk operations that don't need to be atomic.

This command commits an open transaction in this connection.

Example:

SET UNDO_LOG 0

SET WRITE_DELAY

SET WRITE_DELAY int

Set the maximum delay between a commit and flushing the log, in milliseconds. This setting is persistent. The default is 500 ms.

Admin rights are required to execute this command, as it affects all connections. This command commits an open transaction in this connection. This setting can be appended to the database URL: jdbc:h2:test;WRITE_DELAY=0

Example:

SET WRITE_DELAY 2000

SHUTDOWN

SHUTDOWN [ IMMEDIATELY | COMPACT | DEFRAG ]

This statement closes all open connections to the database and closes the database. This command is usually not required, as the database is closed automatically when the last connection to it is closed.

If no option is used, then the database is closed normally. All connections are closed, open transactions are rolled back.

SHUTDOWN COMPACT fully compacts the database (re-creating the database may further reduce the database size). If the database is closed normally (using SHUTDOWN or by closing all connections), then the database is also compacted, but only for at most the time defined by the database setting h2.maxCompactTime in milliseconds (see there).

SHUTDOWN IMMEDIATELY closes the database files without any cleanup and without compacting.

SHUTDOWN DEFRAG re-orders the pages when closing the database so that table scans are faster. In case of MVStore it is currently equivalent to COMPACT.

Admin rights are required to execute this command.

Example:

SHUTDOWN COMPACT