Data Definition Language (DDL)

Data Manipulation Language (DML)

Bulk Data Import Commands

Bulk Data Export Commands

Authorisation, Data Access Control Language

Transaction Control Commands

Commands to Analyse / Optimize SQL Execution

Pseudo Columns

Details

CREATE SCHEMA

   CREATE SCHEMA [ IF NOT EXISTS ] schema_name
    [ AUTHORISATION auth_name ]

Define a new schema

Example:

CREATE SCHEMA tst;
SET SCHEMA tst;
SELECT CURRENT_SCHEMA;

Note: The "IF NOT EXISTS" option is supported from release Jul2017 (11.27.1) onwards.

For details see: Documentation/SQLreference/Schema
See also: SET SCHEMA   CURRENT_SCHEMA   COMMENT ON SCHEMA   DROP SCHEMA

Associated system table: sys.schemas
 

ALTER SCHEMA

   ALTER SCHEMA [ IF NOT EXISTS ] schema_name RENAME TO new_schema_name

Change the name of a schema

Example:

CREATE SCHEMA tst;
ALTER SCHEMA tst RENAME TO tst2;

It is only allowed to change the name of a schema if no objects exists which depend on the schema name, such as tables, views, functions, etc.
Note: This command is supported from release Apr2019 onwards.

For details see: Documentation/SQLreference/Schema
See also: SET SCHEMA   CURRENT_SCHEMA   COMMENT ON SCHEMA   DROP SCHEMA

Associated system table: sys.schemas
 

CREATE SEQUENCE

   CREATE SEQUENCE [ schema_name . ] sequence_name
    [ AS datatype ]
    [ START WITH bigint# ]
    [ INCREMENT BY bigint# ]
    [ MINVALUE bigint# | NO MINVALUE ]
    [ MAXVALUE bigint# | NO MAXVALUE ]
    [ CACHE bigint# ]
    [ [ NO ] CYCLE ]

Define a new integer number sequence generator

For details see: Documentation/Manuals/SQLreference/SerialTypes
See also: ALTER SEQUENCE   COMMENT ON SEQUENCE   NEXT VALUE FOR   DROP SEQUENCE

Associated system table: sys.sequences
 

CREATE TABLE

   CREATE TABLE [ IF NOT EXISTS ] [ schema_name . ] table_name
    ( column_definition(s)_and_optional_table-constraints_list )

Define a new table including data integrity constraints

Note: WARNING: Column CHECK constraint definitions are accepted but not enforced! They are also not stored in the data dictionary, so will be lost when using msqldump.
The "IF NOT EXISTS" option is supported from release Jul2017 (11.27.1) onwards.
You can also use LIKE qname as part of the column definition to copy the column definitions of qname excluding their constraints. For instance CREATE TABLE webshop.products_new (LIKE webshop.products, descr VARCHAR(9999), pict BLOB);

For details see: Documentation/Manuals/SQLreference/Tables and: Documentation/SQLreference/TableIdentityColumn
See also: ALTER TABLE ADD COLUMN   ALTER TABLE ADD CONSTRAINT   COMMENT ON TABLE   COMMENT ON COLUMN   DROP TABLE

Associated system table: sys.tables where type = 0
 

CREATE TABLE AS

   CREATE TABLE [ IF NOT EXISTS ] [ schema_name . ] table_name
    [ ( column_name [, column_name] [, ... ] ) ]
    AS SELECT_query
    [ WITH [ NO ] DATA ]

Define a new table from the results of a query. By default the table will be populated with the data of the query. Specify WITH NO DATA to only create the table.

Note: Default behavior is WITH DATA.
The "IF NOT EXISTS" option is supported from release Jul2017 (11.27.1) onwards.

For details see: Documentation/Manuals/SQLreference/Tables
See also: ALTER TABLE ADD COLUMN   ALTER TABLE ADD CONSTRAINT   COMMENT ON TABLE   COMMENT ON COLUMN   DROP TABLE

Associated system table: sys.tables where type = 0
 

CREATE TABLE FROM LOADER

   CREATE TABLE [ IF NOT EXISTS ] [ schema_name . ] table_name
    FROM LOADER function_name ( [ arg1_val [ , arg2_val ] [, ... ] ] )
    [ WITH [ NO ] DATA ]

Define a new table from the results of a (Python) loader function. By default the table will be populated with the data of the loader function. Specify WITH NO DATA to only create the table.

Note: This command is MonetDB specific. It is introduced in release Dec2016 (11.25.3).
The "IF NOT EXISTS" option is supported from release Jul2017 (11.27.1) onwards.

For details see: blog/monetdbpython-loader-functions
See also: CREATE LOADER   COPY LOADER INTO FROM      COMMENT ON TABLE   COMMENT ON COLUMN   DROP TABLE

Associated system table: sys.tables where type = 0
 

CREATE MERGE TABLE

   CREATE MERGE TABLE [ IF NOT EXISTS ] [ schema_name . ] table_name
    ( column_definition(s)_and_optional_table-constraints_list )
    [ PARTITION BY { RANGE | VALUES }  { ON ( column_name )  |  USING ( expression ) } ]

Define a new merge table to create a horizontally partitioned table.
A merge table logically combines the data of multiple other tables (called partition tables which are added via ALTER TABLE merge_table ADD TABLE partition_table) which all must have the exact same table definition. This merge table is easier to extend/change with new partitions than a view which combines multiple SELECT queries (via UNION ALL) and can processes queries faster. Also with "PARTITION BY" specified, the virtual merge table becomes updatable, so allow inserts, updates, deletes and truncate on the merge table directly instead of the partition tables.

Note: This command is MonetDB specific.
The "PARTITION BY" option is supported from release Apr2019 (11.33.3) onwards. See updatable-merge-tables for details.
The "IF NOT EXISTS" option is supported from release Jul2017 (11.27.1) onwards.
A merge table is removed (including all its partition table information) using the ordinary DROP TABLE statement. There is no DROP MERGE TABLE statement.

For details see: Documentation/Cookbooks/SQLrecipes/DataPartitioning and updatable-merge-tables
and: Documentation/Manuals/SQLreference/Tables
See also: ALTER TABLE ADD TABLE   ALTER TABLE DROP TABLE   COMMENT ON TABLE   COMMENT ON COLUMN   DROP TABLE

Associated system tables: sys.tables where type = 3;   sys.table_partitions
 

CREATE REMOTE TABLE

   CREATE REMOTE TABLE [ IF NOT EXISTS ] [ schema_name . ] table_name
    ( column definition(s) )
    ON remote_table URL
    [ WITH [USER 'user_login_name_nm']   [ [ENCRYPTED] PASSWORD 'password'] ]

Define a alias for a new remote table. The remote table must be an existing table on a running MonetDB server

Note: This command is MonetDB specific.
The "IF NOT EXISTS" option is supported from release Jul2017 (11.27.1) onwards.
The "WITH USER ... PASSWORD ..." option is supported from release Aug2018 (11.31.7) onwards. See this blog post for more information.
A remote table is removed using the ordinary DROP TABLE statement. There is no DROP REMOTE TABLE statement.

For details see: Documentation/Cookbooks/SQLrecipes/DistributedQueryProcessing and: Documentation/Manuals/SQLreference/Tables
See also: COMMENT ON TABLE   COMMENT ON COLUMN   DROP TABLE

Associated system table: sys.tables where type = 5
 

CREATE REPLICA TABLE

   CREATE REPLICA TABLE [ IF NOT EXISTS ] [ schema_name . ] table_name
    ( column_definition(s)_and_optional_table-constraints_list )

Define a new replica table

Note: This command is MonetDB specific.
The "IF NOT EXISTS" option is supported from release Jul2017 (11.27.1) onwards.
A replica table is removed using the ordinary DROP TABLE statement. There is no DROP REPLICA TABLE statement.

For details see: Documentation/Manuals/SQLreference/Tables and: Documentation/SQLreference/TableIdentityColumn
See also: COMMENT ON TABLE   COMMENT ON COLUMN   DROP TABLE

Associated system table: sys.tables where type = 6
 

CREATE STREAM TABLE

   CREATE STREAM TABLE [ IF NOT EXISTS ] [ schema_name . ] table_name
    ( column_definition(s)_and_optional_table-constraints_list )

Define a new stream table

Example:

CREATE STREAM TABLE room_event (ts timestamp, room integer, payload decimal(8,2));

Note: This command is MonetDB specific.
The "IF NOT EXISTS" option is supported from release Jul2017 (11.27.1) onwards.
A stream table is removed using the ordinary DROP TABLE statement. There is no DROP STREAM TABLE statement.

For details see: Documentation/Extensions/Streams/Catalog
See also: COMMENT ON TABLE   COMMENT ON COLUMN   DROP TABLE

Associated system table: sys.tables where type = 4
 

CREATE TEMPORARY TABLE

   CREATE [ LOCAL | GLOBAL ] { TEMPORARY | TEMP } TABLE [ IF NOT EXISTS ] [ schema_name . ] table_name
    { ( column_definition(s)_and_optional_table-constraints_list )  |
      [ ( column_name [, column_name ] [, ... ] ) ] AS SELECT_query [ WITH [ NO ] DATA ] }
    [ ON COMMIT { DELETE ROWS | PRESERVE ROWS | DROP } ]

Define a new temporary table. The visibility of the table to other session users can be controlled by using GLOBAL. Default is LOCAL. A temporary table will be removed automatically after the user session is terminated.
When ON COMMIT clause is not specified then the default behavior is ON COMMIT DELETE ROWS, complying to the SQL standard.
When using AS SELECT ... the default is WITH DATA.

Example:

CREATE TEMP TABLE names (id int NOT NULL PRIMARY KEY, name VARCHAR(99) NOT NULL UNIQUE) ON COMMIT PRESERVE ROWS;
-- Note that temporary tables are implicitly assigned to schema: tmp
INSERT INTO tmp.names VALUES (1, 'one');
INSERT INTO tmp.names VALUES (2, 'two');
INSERT INTO tmp.names VALUES (2, 'dos');
-- Error: INSERT INTO: PRIMARY KEY constraint 'names.names_id_pkey' violated
INSERT INTO tmp.names VALUES (3, 'two');
-- Error: INSERT INTO: UNIQUE constraint 'names.names_name_unique' violated
INSERT INTO tmp.names VALUES (3, 'free');
SELECT * FROM tmp.names;
-- shows 3 rows
DROP TABLE tmp.names;

CREATE GLOBAL TEMP TABLE tmp.name_lengths (name, length)
 AS SELECT DISTINCT name, LENGTH(name) FROM sys.ids ORDER BY 1
 WITH DATA
 ON COMMIT PRESERVE ROWS;
SELECT * FROM tmp.name_lengths WHERE name ILIKE '%\\_id%' ESCAPE '\\';
SELECT COUNT(*) AS count_names, AVG(length) AS avg_name_length FROM tmp.name_lengths;
DROP TABLE tmp.name_lengths;

IMPORTANT: specify ON COMMIT PRESERVE ROWS to keep rows when working in autocommit mode, because default behavior is ON COMMIT DELETE ROWS.
Note: The "IF NOT EXISTS" option is supported from release Jul2017 (11.27.1) onwards.
It is not possible to add comments on temporary tables or columns thereof.
A temporary table is removed using the ordinary DROP TABLE statement. There is no DROP TEMPORARY TABLE statement.

For details see: Documentation/Manuals/SQLreference/Tables and: Documentation/SQLreference/TableIdentityColumn
See also: ALTER TABLE ADD COLUMN   ALTER TABLE ADD CONSTRAINT   DROP TABLE

Associated system table: sys.tables where type in (20, 30)
 

CREATE INDEX

   CREATE [ UNIQUE ] INDEX index_name ON [ schema_name . ] table_name ( column_name [, column_name ] [, ... ] )

Define a new secondary index on one or more columns of a specific table

Note: Although CREATE INDEX commands are accepted by the MonetDB parser for SQL compliance purposes, it currently does not create a physical secondary index via this SQL command. Instead MonetDB internally decides which column search accelerator(s) to create, persist and use during SQL query execution.
The index name must be unique within the schema of the table. You can not specify the schema for the index. It will be placed in the same schema as the table.
WARNING: the UNIQUE keyword does NOT enforce a uniqueness constraint. To create a unique constraint use: ALTER TABLE s.t ADD CONSTRAINT t_uc UNIQUE (c1, c2) instead.

For details see: Documentation/Manuals/SQLreference/Indices
See also: CREATE ORDERED INDEX   ALTER TABLE ADD CONSTRAINT   COMMENT ON INDEX   DROP INDEX

Associated system table: sys.idxs
 

CREATE IMPRINTS INDEX

   CREATE IMPRINTS INDEX index_name ON [ schema_name . ] table_name ( column_name )

Define a new imprints index (a simple but efficient cache conscious secondary index) on one numerical column of a specific table. This index is a special single column index (for numeric columns only) which stores meta data (min, max, nulls) on segments of column data. An imprint is used during query evaluation to limit data access and thus minimize memory traffic. It can speedup queries which have a column selection condition (such as: AGE IS NULL OR AGE BETWEEN 25 AND 65). The storage overhead is just a few percent over the size of the columns being indexed.

Example:

CREATE IMPRINTS INDEX my_impr_idx ON myschema.mytable ( my_num_column );

Note: This command is MonetDB specific. It is introduced in release Dec2016 (11.25.3).
The imprints index is a new type of experimental column index.
Limitations are: Only 1 column can be indexed per index. Only columns of fixed size data type (so not on: char, varchar, clob, blob, url, json, inet and uuid) can be indexed.
The index name must be unique within the schema of the table. You can not specify the schema for the index. It will be placed in the same schema as the table.
Warning: The index is not maintained automatically and will become inactive after inserts, deletes or updates are done on the column data.
An imprints index is removed using the ordinary DROP INDEX statement. There is no DROP IMPRINTS INDEX statement.

For details see: Documentation/Manuals/SQLreference/Indices
and: Home/ScienceLibrary
See also: CREATE ORDERED INDEX   CREATE INDEX   COMMENT ON INDEX   DROP INDEX

Associated system table: sys.idxs   sys.storage
 

CREATE ORDERED INDEX

   CREATE ORDERED INDEX index_name ON [ schema_name . ] table_name ( column_name )

Define a new ordered index on one column of a specific table. This index is a special single column index where the values are stored in ascending order. It can speedup queries which have a column selection condition (such as: AGE >=18 or AGE BETWEEN 18 AND 30) or need sorting such as when GROUP BY is used.

Example:

CREATE ORDERED INDEX my_ord_idx ON myschema.mytable ( my_column );

Note: This command is MonetDB specific. It is introduced in release Dec2016 (11.25.3).
The ordered index is a new type of experimental column index.
Limitations: Only 1 column can be indexed per index.
The index name must be unique within the schema of the table. You can not specify the schema for the index. It will be placed in the same schema as the table.
Releases before Mar2018 (11.29.3) only allowed columns of fixed size data type (so not on: char, varchar, clob, blob, url, json, inet and uuid) could be indexed. This limitation has been removed in Mar2018 (and future) releases.
Warning: The index is not maintained automatically and will become inactive after inserts, deletes or updates are done on the column data.
A ordered index is removed using the ordinary DROP INDEX statement. There is no DROP ORDERED INDEX statement.

For details see: Documentation/Manuals/SQLreference/Indices
See also: CREATE IMPRINTS INDEX   CREATE INDEX   COMMENT ON INDEX   DROP INDEX

Associated system table: sys.idxs   sys.storage
 

CREATE VIEW

   CREATE [ OR REPLACE ] VIEW [ schema_name . ] view_name
    [ ( column_name [, column_name ] [, ... ] ) ]
    AS SELECT_query
    [ WITH CHECK OPTION ]

Define a new SQL view. A view can be seen as a stored SELECT query. It is useful to reduce query complexity as it can include joins, unions and aggregations, increase standardisation and reusability. It can also be used to simplify reporting.

Limitation: Views are not updatable. The "WITH CHECK OPTION" is accepted for compliance but has no effect.
Recursive views and reference-able views are not supported.
Note: An "ORDER BY" clause is supported from release Jul2017 (11.27.1) onwards.
The "OR REPLACE" option is supported from release Mar2018 (11.29.3) onwards.

For details see: Documentation/Manuals/SQLreference/Views
See also: COMMENT ON VIEW   COMMENT ON COLUMN   DROP VIEW

Associated system table: sys.tables where type = 1

CREATE AGGREGATE EXTERNAL

   CREATE [ OR REPLACE ] AGGREGATE [ schema_name . ] aggregate_name ( [ arg1_nm_type [ , arg2_nm_type ] [, ... ] ] )
    RETURNS datatype
    EXTERNAL NAME  MAL_function_name

Define a new user-defined aggregation function for which the implementation is done externally

Example:

CREATE AGGREGATE quantile(val bigint, q double) RETURNS bigint EXTERNAL NAME "aggr"."quantile";

Note: This command is MonetDB specific. External implies language MAL.
To allow other users to invoke a user-defined aggregate function, you must grant the other users (or PUBLIC) EXECUTE privilege for the aggregate function.
The "OR REPLACE" option is supported from release Jul2017 (11.27.1) onwards.

For details see: Documentation/Manuals/SQLreference/Functions
See also: COMMENT ON AGGREGATE   GRANT PRIVILEGES   DROP AGGREGATE   DROP ALL AGGREGATE

Associated system table: sys.functions where type = 3 and language = 1
 

CREATE AGGREGATE LANGUAGE

   CREATE [ OR REPLACE ] AGGREGATE [ schema_name . ] aggregate_name ( [ arg1_nm_type [ , arg2_nm_type ] [, ... ] ] )
    RETURNS datatype
    LANGUAGE { C | CPP | R | PYTHON | PYTHON_MAP | PYTHON2 | PYTHON2_MAP | PYTHON3 | PYTHON3_MAP }
    '{' function_body '}'

Define a new user-defined aggregation function for which the body implementation is specified in the specific script language

Example:

CREATE AGGREGATE python_aggregate(val INTEGER) 
RETURNS INTEGER 
LANGUAGE PYTHON {
    try:
        unique = numpy.unique(aggr_group)
        x = numpy.zeros(shape=(unique.size))
        for i in range(0, unique.size):
            x[i] = numpy.sum(val[aggr_group==unique[i]])
    except NameError:
        # aggr_group doesn't exist. no groups, aggregate on all data
        x = numpy.sum(val)
    return(x)
};

Note: This command is MonetDB specific. Supported languages are: C, C++, R and Python.
If your Python code requires Python 2 to work correctly use PYTHON2 instead of PYTHON. If your Python code requires Python 3 to work correctly use PYTHON3 instead of PYTHON. If your Python code can be executed in parallel (using multiple threads) without side effects, use PYTHON_MAP instead of PYTHON. Idem for PYTHON2_MAP and PYTHON3_MAP.
For languages C and CPP a C/C++ compiler must be available on the deployment server and the MonetDB server started with option: --set embedded_c=true.
For language R the R script interpreter software must be available on the deployment server and the MonetDB server started with option: --set embedded_r=true.
For language PYTHON the Python script interpreter software must be available on the deployment server and the MonetDB server started with option: --set embedded_py=true  or --set embedded_py=2  or --set embedded_py=3.
To allow other users to invoke a user-defined aggregate function, you must grant the other users (or PUBLIC) EXECUTE privilege for the aggregate function.
The "OR REPLACE" option is supported from release Jul2017 (11.27.1) onwards.

For details see: JIT C/C++ UDFs and embedded-pythonnumpy-monetdb and embedded-r-monetdb
See also: COMMENT ON AGGREGATE   GRANT PRIVILEGES   DROP AGGREGATE   CREATE FUNCTION LANGUAGE

Associated system table: sys.functions where type = 3 and language > 2; sys.function_languages
 

CREATE FUNCTION

   CREATE [ OR REPLACE ] FUNCTION [ schema_name . ] function_name ( [ arg1_nm_type [ , arg2_nm_type ] [, ... ] ] )
    RETURNS datatype
    BEGIN function_body END

Define a new user-defined function

Example:

CREATE FUNCTION heapspace(tpe string, i bigint, w int) returns bigint
begin
 if tpe <> 'varchar' and tpe <> 'clob' then return 0;
 end if;
 return 10240 + i * w;
end;

Note: Besides returning a scalar value (type = 1), functions can also be defined to return a Table as data type (type = 5).
To allow other users to invoke a user-defined function, you must grant the other users (or PUBLIC) EXECUTE privilege for the function.
The "OR REPLACE" option is supported from release Jul2017 (11.27.1) onwards.

For details see: Documentation/Manuals/SQLreference/Functions
See also: COMMENT ON FUNCTION   GRANT PRIVILEGES   DROP FUNCTION   DROP ALL FUNCTION

Associated system table: sys.functions where type in (1,5) and language = 2
 

CREATE FUNCTION EXTERNAL

   CREATE [ OR REPLACE ] FUNCTION [ schema_name . ] function_name ( [ arg1_nm_type [ , arg2_nm_type ] [, ... ] ] )
    RETURNS datatype
    EXTERNAL NAME  MAL_function_name

Define a new user-defined function for which the implementation is done externally

Example:

CREATE FUNCTION isa_uuid(s string) returns boolean external name uuid."isaUUID";

Note: This command is MonetDB specific. External implies language MAL.
Besides returning a scalar value (type = 1), functions can also be defined to return a Table as data type (type = 5).
To allow other users to invoke a user-defined function, you must grant the other users (or PUBLIC) EXECUTE privilege for the function.
The "OR REPLACE" option is supported from release Jul2017 (11.27.1) onwards.

For details see: Documentation/Manuals/SQLreference/Functions
and: Documentation/Cookbooks/SQLrecipes/UserDefinedFunction
See also: COMMENT ON FUNCTION   GRANT PRIVILEGES   DROP FUNCTION   CREATE FUNCTION   CREATE FUNCTION LANGUAGE

Associated system table: sys.functions where type in (1,5) and language = 1
 

CREATE FUNCTION LANGUAGE

   CREATE [ OR REPLACE ] FUNCTION [ schema_name . ] function_name ( [ arg1_nm_type [ , arg2_nm_type ] [, ... ] ] )
    RETURNS datatype
    LANGUAGE { C | CPP | R | PYTHON | PYTHON_MAP | PYTHON2 | PYTHON2_MAP | PYTHON3 | PYTHON3_MAP }
    '{' function_body '}'

Define a new user-defined function for which the implementation is specified in a specific script language

Note: This command is MonetDB specific. Supported languages are: C, C++, R and Python.
If your Python code requires Python 2 to work correctly use PYTHON2 instead of PYTHON. If your Python code requires Python 3 to work correctly use PYTHON3 instead of PYTHON. If your Python code can be executed in parallel (using multiple threads) without side effects, use PYTHON_MAP instead of PYTHON. Idem for PYTHON2_MAP and PYTHON3_MAP.
For languages C and CPP a C/C++ compiler must be available on the deployment server and the MonetDB server started with option: --set embedded_c=true.
For language R the R script interpreter software must be available on the deployment server and the MonetDB server started with option: --set embedded_r=true.
For language PYTHON the Python script interpreter software must be available on the deployment server and the MonetDB server started with option: --set embedded_py=true or --set embedded_py=2 or --set embedded_py=3.
Besides returning a scalar value (type = 1), functions can also be defined to return a Table as data type (type = 5).
To allow other users to invoke a user-defined function, you must grant the other users (or PUBLIC) EXECUTE privilege for the function.
The "OR REPLACE" option is supported from release Jul2017 (11.27.1) onwards.

For details see: JIT C/C++ UDFs and embedded-pythonnumpy-monetdb and embedded-r-monetdb and voter-classification-using-monetdbpython
See also: COMMENT ON FUNCTION   GRANT PRIVILEGES   DROP FUNCTION   CREATE AGGREGATE LANGUAGE   CREATE LOADER

Associated system table: sys.functions where type in (1,5) and language > 2; sys.function_languages
 

CREATE FILTER FUNCTION EXTERNAL

   CREATE [ OR REPLACE ] FILTER FUNCTION [ schema_name . ] function_name ( [ arg1_nm_type [ , arg2_nm_type ] [, ... ] ] )
    RETURNS datatype
    EXTERNAL NAME  MAL_function_name

Define a new user-defined filter function for which the implementation is done externally

Example:

CREATE FILTER FUNCTION "re_like"(val string, pat string) external name algebra."rexpr_like";

Note: This command is MonetDB specific. External implies language MAL.
To allow other users to invoke a user-defined filter function, you must grant the other users (or PUBLIC) EXECUTE privilege for the filter function.
The "OR REPLACE" option is supported from release Jul2017 (11.27.1) onwards.

See also: COMMENT ON FILTER FUNCTION   GRANT PRIVILEGES   DROP FILTER FUNCTION   DROP ALL FILTER FUNCTION

Associated system table: sys.functions where type = 4 and language = 1
 

CREATE LOADER

   CREATE [ OR REPLACE ] LOADER [ schema_name . ] function_name ( [ arg1_nm_type [ , arg2_nm_type ] [, ... ] ] )
    LANGUAGE PYTHON
    '{' python_code_with_emit.emit()_function '}'

Define a new user-defined loader function for which the implementation is done in Python language. A loader function can for instance read data from an external file is a specific format such as XML, json, bson, ods, xlsx, etc.

Example:

CREATE LOADER json_loader(filename STRING) LANGUAGE PYTHON {
    import json
    f = open(filename)
    _emit.emit(json.load(f))
    f.close()
};

Note: This command is MonetDB specific. It is introduced in release Dec2016 (11.25.3).
For language PYTHON the Python script interpreter software must be available on the deployment server and the MonetDB server started with option: --set embedded_py=true  or --set embedded_py=2  or --set embedded_py=3.
To allow other users to invoke a user-defined loader function, you must grant the other users (or PUBLIC) EXECUTE privilege for the loader function.
The "OR REPLACE" option is supported from release Jul2017 (11.27.1) onwards.

For details see: blog/monetdbpython-loader-functions
See also: COMMENT ON LOADER   GRANT PRIVILEGES   DROP LOADER   CREATE TABLE FROM LOADER   COPY LOADER INTO FROM

Associated system table: sys.functions where type = 7 and language > 2
 

CREATE PROCEDURE

   CREATE [ OR REPLACE ] PROCEDURE [ schema_name . ] procedure_name ( [ arg1_nm_type [ , arg2_nm_type ] [, ... ] ] )
    BEGIN procedure_body END

Define a new user-defined procedure

Note: To allow other users to call and execute a user-defined procedure, you must grant the other users (or PUBLIC) EXECUTE privilege for the procedure.
The "OR REPLACE" option is supported from release Jul2017 (11.27.1) onwards.

For details see: Documentation/Manuals/SQLreference/Procedures
See also: COMMENT ON PROCEDURE   DECLARE   CALL   GRANT PRIVILEGES   DROP PROCEDURE   DROP ALL PROCEDURE

Associated system table: sys.functions where type = 2 and language = 2
 

CREATE PROCEDURE EXTERNAL

   CREATE [ OR REPLACE ] PROCEDURE [ schema_name . ] procedure_name ( [ arg1_nm_type [ , arg2_nm_type ] [, ... ] ] )
    EXTERNAL NAME  MAL_procedure_name

Define a new user-defined procedure for which the implementation is done externally

Example:

CREATE PROCEDURE sys.createorderindex(sys string, tab string, col string) external name sql.createorderindex;

Note: This command is MonetDB specific. External implies language MAL.
To allow other users to call and execute a user-defined procedure, you must grant the other users (or PUBLIC) EXECUTE privilege for the procedure.
The "OR REPLACE" option is supported from release Jul2017 (11.27.1) onwards.

For details see: Documentation/Manuals/SQLreference/Procedures
See also: COMMENT ON PROCEDURE   CALL   GRANT PRIVILEGES   DROP PROCEDURE   DROP ALL PROCEDURE

Associated system table: sys.functions where type = 2 and language = 1
 

CREATE TYPE EXTERNAL

   CREATE TYPE [ schema_name . ] type_name
    EXTERNAL NAME  MAL type_name

Declare a new user-defined data type. The implementation (structure, operators & functions both scalar and bulk) must be done externally in compiled C code and MAL script. For examples see the C implementation of types: inet, json, url and uuid.

Note: This command is MonetDB specific. External implies language MAL.

For details see: Documentation/Manuals/SQLreference/Userdefinedtypes
See also: DROP TYPE   CREATE TABLE

Associated system table: sys.types where eclass = 16

CREATE TRIGGER

   CREATE [ OR REPLACE ] TRIGGER [ schema_name . ] trigger_name
    { BEFORE | AFTER }
    { INSERT | DELETE | TRUNCATE | UPDATE [ OF column_name [, column_name] [, ... ] ] }
    ON [ schema_name . ] table_name
    [ REFERENCING { { OLD | NEW } { [ ROW ] | TABLE } [ AS ] ident } [...] ]
    [ FOR [ EACH ] { ROW | STATEMENT } ]
    [ WHEN ( search_condition ) ]
    { trigger_procedure_statement | BEGIN ATOMIC trigger_procedure_statement_list END }

Define a new trigger on a table update event

Note: FOR EACH STATEMENT is the default if not specified.
The schema name of a full qualified trigger name must be the same as the schema name of the table.
The "OR REPLACE" option and the TRUNCATE trigger event are supported from release Mar2018 (11.29.3) onwards.

For details see: Documentation/SQLreference/Triggers
See also: DROP TRIGGER   DECLARE   GRANT PRIVILEGES

Associated system table: sys.triggers
 

COMMENT ON SCHEMA

   COMMENT ON SCHEMA schema_name IS  { 'comment text' | NULL | '' }

Add or update or remove a comment for a schema

Example:

COMMENT ON SCHEMA prj4 IS 'schema of new project 4';
COMMENT ON SCHEMA prj0 IS '';
COMMENT ON SCHEMA prjX IS NULL;

Note: By specifying IS NULL or IS '' you remove the comment for the schema.
If the schema is dropped, the associated comment is also removed.
This command is supported from release Mar2018 (11.29.3) onwards.

For details see: bugzilla/show_bug.cgi?id=6109#c5
See also: CREATE SCHEMA   DROP SCHEMA

Associated system table: sys.comments
 

COMMENT ON TABLE

   COMMENT ON TABLE [ schema_name . ] table_name IS  { 'comment text' | NULL | '' }

Add or update or remove a comment for a table

Example:

COMMENT ON TABLE sys.comments IS 'contains comments on all db objects'

Note: By specifying IS NULL or IS '' you remove the comment for the table object.
If the table is dropped, the associated comment (including the comments for the columns) is also removed.
This command is supported from release Mar2018 (11.29.3) onwards.
Note: it is not allowed or possible to add comments for temporary tables or objects in schema "tmp".

For details see: bugzilla/show_bug.cgi?id=6109#c5
See also: CREATE TABLE   DROP TABLE

Associated system table: sys.comments
 

COMMENT ON VIEW

   COMMENT ON VIEW [ schema_name . ] view_name IS  { 'comment text' | NULL | '' }

Add or update or remove a comment for a view

Example:

COMMENT ON VIEW mysch.articles_aggr IS 'view lists aggregated info on articles'

Note: By specifying IS NULL or IS '' you remove the comment for the view.
If the view is dropped, the associated comment (including the comments for the columns) is also removed.
This command is supported from release Mar2018 (11.29.3) onwards.

For details see: bugzilla/show_bug.cgi?id=6109#c5
See also: CREATE VIEW   DROP VIEW

Associated system table: sys.comments
 

COMMENT ON COLUMN

   COMMENT ON COLUMN [ schema_name . ] table_or_view_name . column_name IS  { 'comment text' | NULL | '' }

Add or update or remove a comment for a column of a table or view

Example:

COMMENT ON COLUMN sys.comments.remark IS 'contains description text'

Note: By specifying IS NULL or IS '' you remove the comment for the column.
If the column (or table or view) is dropped, the associated comment is also removed.
This command is supported from release Mar2018 (11.29.3) onwards.
Note: it is not allowed or possible to add comments for columns of temporary tables or objects in schema "tmp".

For details see: bugzilla/show_bug.cgi?id=6109#c5
See also: ALTER TABLE ADD COLUMN   ALTER TABLE DROP COLUMN

Associated system table: sys.comments
 

COMMENT ON INDEX

   COMMENT ON INDEX [ schema_name . ] index_name IS  { 'comment text' | NULL | '' }

Add or update or remove a comment for an index

Example:

COMMENT ON INDEX mysch.article_id IS 'unique index of id key column of article table'

Note: By specifying IS NULL or IS '' you remove the comment for the index.
If the index is dropped, the associated comment is also removed.
This command is supported from release Mar2018 (11.29.3) onwards.

For details see: bugzilla/show_bug.cgi?id=6109#c5
See also: CREATE INDEX   DROP INDEX

Associated system table: sys.comments
 

COMMENT ON SEQUENCE

   COMMENT ON SEQUENCE [ schema_name . ] sequence_name IS  { 'comment text' | NULL | '' }

Add or update or remove a comment for a sequence

Example:

COMMENT ON SEQUENCE mysch.article_id_seq IS 'sequence for article id column'

Note: By specifying IS NULL or IS '' you remove the comment for the sequence.
If the sequence is dropped, the associated comment is also removed.
This command is supported from release Mar2018 (11.29.3) onwards.

For details see: bugzilla/show_bug.cgi?id=6109#c5
See also: CREATE SEQUENCE   DROP SEQUENCE

Associated system table: sys.comments
 

COMMENT ON FUNCTION

   COMMENT ON FUNCTION [ schema_name . ] function_name IS  { 'comment text' | NULL | '' }

Add or update or remove a comment for a function

Example:

COMMENT ON FUNCTION sys.upper IS 'converts text into uppercase'

Note: By specifying IS NULL or IS '' you remove the comment for the function.
If the function is dropped, the associated comment is also removed.
This command is supported from release Mar2018 (11.29.3) onwards.

For details see: bugzilla/show_bug.cgi?id=6109#c5
See also: CREATE FUNCTION   DROP FUNCTION

Associated system table: sys.comments
 

COMMENT ON PROCEDURE

   COMMENT ON PROCEDURE [ schema_name . ] procedure_name IS  { 'comment text' | NULL | '' }

Add or update or remove a comment for a procedure

Example:

COMMENT ON PROCEDURE mysch.load_articles IS 'proc which reloads the articles from external file article.csv'

Note: By specifying IS NULL or IS '' you remove the comment for the procedure.
If the procedure is dropped, the associated comment is also removed.
This command is supported from release Mar2018 (11.29.3) onwards.

For details see: bugzilla/show_bug.cgi?id=6109#c5
See also: CREATE PROCEDURE   DROP PROCEDURE

Associated system table: sys.comments
 

COMMENT ON AGGREGATE

   COMMENT ON AGGREGATE [ schema_name . ] aggregate_name IS  { 'comment text' | NULL | '' }

Add or update or remove a comment for an aggregate function

Example:

COMMENT ON AGGREGATE sys.std_dev IS 'computes the standard deviation of a group of numeric values'

Note: By specifying IS NULL or IS '' you remove the comment for the aggregate function.
If the aggregate function is dropped, the associated comment is also removed.
This command is supported from release Mar2018 (11.29.3) onwards.

For details see: bugzilla/show_bug.cgi?id=6109#c5
See also: CREATE AGGREGATE LANGUAGE   DROP AGGREGATE

Associated system table: sys.comments
 

COMMENT ON FILTER FUNCTION

   COMMENT ON FILTER FUNCTION [ schema_name . ] function_name IS  { 'comment text' | NULL | '' }

Add or update or remove a comment for a filter function

Example:

COMMENT ON FILTER FUNCTION sys."ilike"(clob, clob) IS 'case insensitive pattern matching';
COMMENT ON FILTER FUNCTION sys."ilike"(clob, clob, clob) IS 'case insensitive pattern matching with user specified escape character';

Note: By specifying IS NULL or IS '' you remove the comment for the filter function.
If the filter function is dropped, the associated comment is also removed.
This command is supported from release Mar2018 (11.29.3) onwards.

For details see: bugzilla/show_bug.cgi?id=6109#c5
See also: CREATE FILTER FUNCTION EXTERNAL   DROP FILTER FUNCTION

Associated system table: sys.comments
 

COMMENT ON LOADER

   COMMENT ON LOADER [ schema_name . ] function_name IS  { 'comment text' | NULL | '' }

Add or update or remove a comment for a loader function

Example:

COMMENT ON LOADER mysch.load_xls_data IS 'custom loader to extract data from external xls file'

Note: By specifying IS NULL or IS '' you remove the comment for the loader function.
If the loader function is dropped, the associated comment is also removed.
This command is supported from release Mar2018 (11.29.3) onwards.

For details see: bugzilla/show_bug.cgi?id=6109#c5
See also: CREATE LOADER   DROP LOADER

Associated system table: sys.comments
 

DECLARE

   DECLARE variable_name datatype

Example:

DECLARE ts1 timestamp;
SET ts1 = now();
SELECT ts1;
SELECT * FROM sys.var() WHERE name NOT IN (SELECT var_name FROM sys.var_values);

Note: The declared variable is not persistent. It will be lost after closing the connection or session.

For details see: Documentation/Manuals/SQLreference/Variables
See also: SET   SELECT

Associated system table: sys.var()
 

ALTER SEQUENCE

   ALTER SEQUENCE [ schema_name . ] sequence_name [ AS datatype  ]
    [ RESTART [WITH bigint# ] ]
    [ INCREMENT BY bigint# ]
    [ MINVALUE bigint# | NO MINVALUE ]
    [ MAXVALUE bigint# | NO MAXVALUE ]
    [ CACHE bigint# ]
    [ [ NO ] CYCLE ]

Change definition of a sequence generator

For details see: Documentation/Manuals/SQLreference/SerialTypes
See also: DROP SEQUENCE   CREATE SEQUENCE

Associated system table: sys.sequences
 

ALTER TABLE ADD COLUMN

   ALTER TABLE [ IF EXISTS ] [ schema_name . ] table_name
    ADD [ COLUMN ] column_name { data_type [ column_option ... ] | SERIAL | BIGSERIAL }

Add a column to a table

Note: The "IF EXISTS" option is supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/SQLreference/Alter
See also: ALTER TABLE ADD CONSTRAINT   ALTER TABLE ALTER COLUMN   ALTER TABLE DROP COLUMN

Associated system table: sys.columns
 

ALTER TABLE ALTER COLUMN

   ALTER TABLE [ IF EXISTS ] [ schema_name . ] table_name
    ALTER [ COLUMN ] column_name { SET NULL | SET NOT NULL | SET DEFAULT value | DROP DEFAULT | SET STORAGE {string | NULL} }

Change column nullability or default value or storage

Note: Change of the data type of a column is not supported. Instead use command sequence:
  ALTER TABLE tbl ADD COLUMN new_column new_data_type;
  UPDATE tbl SET new_column = CONVERT(old_column, new_data_type);
  ALTER TABLE tbl DROP COLUMN old_column RESTRICT;
  ALTER TABLE tbl RENAME COLUMN new_column TO old_column;

The "IF EXISTS" option is supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/SQLreference/Alter
See also: ALTER TABLE RENAME COLUMN   ALTER TABLE ADD COLUMN   ALTER TABLE DROP COLUMN   ALTER TABLE DROP CONSTRAINT

Associated system table: sys.columns
 

ALTER TABLE DROP COLUMN

   ALTER TABLE [ IF EXISTS ] [ schema_name . ] table_name
    DROP [ COLUMN ] column_name
    [ RESTRICT | CASCADE ]

Remove a column from a table

Note: You cannot drop a column if it is referenced (e.g. from a view, an index, a merge table, a trigger, a foreign key constraint, a function or procedure or another db object except comments). Use option CASCADE to specify to also drop those referencing objects.
The "IF EXISTS" option is supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/SQLreference/Alter
See also: ALTER TABLE DROP CONSTRAINT   ALTER TABLE ALTER COLUMN   DROP TABLE

Associated system table: sys.columns
 

ALTER TABLE ADD CONSTRAINT

   ALTER TABLE [ IF EXISTS ] [ schema_name . ] table_name
    ADD [ CONSTRAINT constraint_name ]
    { PRIMARY KEY ( column_name [ , column_name ] [, ... ] )
    |      UNIQUE ( column_name [ , column_name ] [, ... ] )
    | FOREIGN KEY ( column_name [ , column_name ] [, ... ] ) REFERENCES [ schema_name . ] table_name
        [ ( column_name [ , column_name ] [, ... ] ) ]   [ match_options ]   [ ref_actions ]
    }

Add a table constraint to a table

Examples:

ALTER TABLE "tblnm" ADD PRIMARY KEY ("C1_id");
ALTER TABLE if exists "schnm"."tblnm" ADD CONSTRAINT "tblnm_uc" UNIQUE ("name", "desc");
ALTER TABLE "tblnm" ADD CONSTRAINT "tblnm_fk1" FOREIGN KEY ("f_id", "f_seq") REFERENCES "schnm2"."fun" ("id", "seq");

Note: Only one PRIMARY KEY constraint can be defined per table. When a primary key constraint is added, all the primary key columns will become NOT NULLable implicitly.
If no constraint_name is specified a constraint_name will be composed implicitly from the table name, column name(s) and constraint type.
The CHECK constraint is not (yet) supported.
The "IF EXISTS" option is supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/SQLreference/Alter
See also: ALTER TABLE DROP CONSTRAINT   ALTER TABLE ALTER COLUMN

Associated system table: sys.keys
 

ALTER TABLE DROP CONSTRAINT

   ALTER TABLE [ IF EXISTS ] [ schema_name . ] table_name
    DROP CONSTRAINT  constraint_name
    [ RESTRICT | CASCADE ]

Remove a table/column constraint from a table

Note: You cannot drop a constraint if it is referenced (e.g. from a foreign key constraint). Use option CASCADE to specify to also drop those referencing objects.
To remove a NOT NULL column constraint use: ALTER TABLE sch.tbl ALTER COLUMN column_name SET NULL.
The "IF EXISTS" option is supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/SQLreference/Alter
See also: ALTER TABLE ADD CONSTRAINT   ALTER TABLE ALTER COLUMN

Associated system table: sys.keys
 

ALTER TABLE RENAME TO

   ALTER TABLE [ IF EXISTS ] [ schema_name . ] table_name
    RENAME TO  new_table_name

Change the name of a table

Note: It is only allowed to change the name of a table if no objects exists which depends on the table name, such as foreign key constraints, views, triggers, indices, functions, procedures, etc.
To move a table to s different schema use command: ALTER TABLE ... SET SCHEMA ...
This command is supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/SQLreference/Alter
See also: CREATE TABLE AS   ALTER TABLE SET SCHEMA

Associated system table: sys.tables
 

ALTER TABLE RENAME COLUMN

   ALTER TABLE [ IF EXISTS ] [ schema_name . ] table_name
    RENAME [ COLUMN ]  column_name TO  new_column_name

Change the name of a column

Note: It is only allowed to change the name of a column if no objects exists which depends on the column name, such as constraints, views, functions, etc.
This command is supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/SQLreference/Alter
See also: ALTER TABLE ADD COLUMN   ALTER TABLE DROP COLUMN   ALTER TABLE DROP CONSTRAINT

Associated system table: sys.columns
 

ALTER TABLE SET SCHEMA

   ALTER TABLE [ IF EXISTS ] [ schema_name . ] table_name
    SET SCHEMA  new_schema_name

Change the schema name of a table

Note: It is only allowed to change the schema name of a table if no objects exists which depends on the table name, such as foreign key constraints, views, triggers, indices, functions, procedures, etc.
This command is supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/SQLreference/Alter
See also: CREATE TABLE AS   ALTER TABLE RENAME TO

Associated system table: sys.tables
 

ALTER TABLE ADD TABLE

   ALTER TABLE [ IF EXISTS ] [ schema_name . ] merge_table_name
    ADD TABLE [ schema_name . ] table_name
    [ AS PARTITION  partition_spec ]

partition_spec:

      IN ( expression [ , expression ] [, ... ] )  [ WITH NULL VALUES ]
    | FROM  { RANGE MINVALUE | expression }  TO  { RANGE MAXVALUE | expression }  [ WITH NULL VALUES ]
    | FOR NULL VALUES

Add a table reference to a merge table set optionally with an partitioning specification

Note: This command is MonetDB specific.
Limitations: The added table must have the same column definitions and layout as the merge table.
Only user defined tables and merge tables can be added to a merge table set.
When the merge table was created with a PARTITION BY clause, the AS PARTITION clause must be specified.
The "AS PARTITION" option is supported from release Apr2019 (11.33.3) onwards.
The "IF EXISTS" option is supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/Cookbooks/SQLrecipes/DataPartitioning and updatable-merge-tables
and: Documentation/SQLreference/Alter
See also: CREATE MERGE TABLE   ALTER TABLE SET TABLE   ALTER TABLE DROP TABLE

Associated system tables: sys.tables where type = 3;   sys.range_partitions;   sys.value_partitions;
 

ALTER TABLE SET TABLE

   ALTER TABLE [ IF EXISTS ] [ schema_name . ] merge_table_name
    SET TABLE [ schema_name . ] table_name
    AS PARTITION  partition_spec

partition_spec:

      IN ( expression [ , expression ] [, ... ] )  [ WITH NULL VALUES ]
    | FROM  { RANGE MINVALUE | expression }  TO  { RANGE MAXVALUE | expression }  [ WITH NULL VALUES ]
    | FOR NULL VALUES

Change the partitioning specification of a partition table

Note: This command is MonetDB specific. It is supported from release Apr2019 (11.33.3) onwards.

For details see: updatable-merge-tables and Documentation/SQLreference/Alter
See also: CREATE MERGE TABLE   ALTER TABLE ADD TABLE   ALTER TABLE DROP TABLE

Associated system tables: sys.tables where type = 3;   sys.range_partitions;   sys.value_partitions;
 

ALTER TABLE DROP TABLE

   ALTER TABLE [ IF EXISTS ] [ schema_name . ] merge_table_name
    DROP TABLE [ schema_name . ] table_name
    [ RESTRICT | CASCADE ]

Remove a table reference from a merge table set. The partition table itself will not be dropped.

Note: This command is MonetDB specific.
The "IF EXISTS" option is supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/Cookbooks/SQLrecipes/DataPartitioning
and: Documentation/SQLreference/Alter
See also: CREATE MERGE TABLE   ALTER TABLE ADD TABLE

Associated system tables: sys.tables where type = 3;   sys.range_partitions;   sys.value_partitions;
 

ALTER TABLE SET INSERT ONLY

   ALTER TABLE [ IF EXISTS ] [ schema_name . ] table_name
    SET INSERT ONLY

Change access of a table to allow only inserts

Note: The "IF EXISTS" option is supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/SQLreference/Alter
See also: ALTER TABLE SET READ ONLY   ALTER TABLE SET READ WRITE

Associated system table: sys.tables where type in (0, 3, 4, 5, 6)
 

ALTER TABLE SET READ ONLY

   ALTER TABLE [ IF EXISTS ] [ schema_name . ] table_name
    SET READ ONLY

Change access of a table to allow only reads (select queries)

Note: The "IF EXISTS" option is supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/SQLreference/Alter
See also: ALTER TABLE SET READ WRITE   ALTER TABLE SET INSERT ONLY

Associated system table: sys.tables
 

ALTER TABLE SET READ WRITE

   ALTER TABLE [ IF EXISTS ] [ schema_name . ] table_name
    SET READ WRITE

Restore access of a table to allow inserts, updates, deletes and reads

Note: The "IF EXISTS" option is supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/SQLreference/Alter
See also: ALTER TABLE SET READ ONLY   ALTER TABLE SET INSERT ONLY

Associated system table: sys.tables where type in (0, 3, 4, 5, 6)
 

DROP SCHEMA

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

Remove a database schema, potentially including all its dependent objects

Example:

DROP SCHEMA IF EXISTS "my_project" RESTRICT;

Note: You cannot drop a schema if it contains objects (e.g. tables, views, functions or procedures, except comments). Use option CASCADE to specify to also drop those referencing objects. The default behavior is RESTRICT.
System schemas (sys, tmp, profiler, json) can not be dropped as they are needed by the system.
Warning: in releases before Mar2018 (11.29.3) the default behavior was CASCADE always. The option RESTRICT was recognised but ignored without warning. This has been corrected in release Mar2018.
The "IF EXISTS" option is supported from release Jul2017 (11.27.1) onwards.

For details see: Documentation/SQLreference/Schema
See also: CREATE SCHEMA   CURRENT_SCHEMA   SET SCHEMA

Associated system table: sys.schemas
 

DROP SEQUENCE

   DROP SEQUENCE [ schema_name . ] sequence_name

Remove an existing sequence generator

Note: You cannot drop a sequence if it is referenced from another object e.g. from a column, a function or procedure.

For details see: Documentation/Manuals/SQLreference/SerialTypes
and: Documentation/Manuals/SQLreference/DropStatement
See also: ALTER SEQUENCE   CREATE SEQUENCE

Associated system table: sys.sequences
 

DROP TABLE

   DROP TABLE [ IF EXISTS ] [ schema_name . ] table_name
    [ RESTRICT | CASCADE ]

Remove a table, potentially including all its dependent objects

Example:

DROP TABLE IF EXISTS "my_project"."my_Table" CASCADE;

Note: You cannot drop a table if it is referenced from a view, a merge table, a trigger, a foreign key constraint, a function, a procedure. Use option CASCADE to specify to also drop those referencing objects. Objects which are part of the table such as columns, pkey/unique/fkey constraints, indices, comments, table_partitions, range_partitions and value_partitions are removed also when a table is dropped.
Warning: When dropping a merge table, also all the related table partitioning information (in sys.table_partitions, sys.range_partitions and sys.value_partitions) is removed.
System tables can also not be dropped as they are needed by the system.
The "IF EXISTS" option is supported from release Jul2017 (11.27.1) onwards.

For details see: Documentation/Cookbooks/SQLrecipes/DataPartitioning
and: Documentation/Manuals/SQLreference/DropStatement
See also: CREATE TABLE

Associated system tables: sys.tables; sys.columns;   sys.comments;   sys.idxs;   sys.table_partitions;   sys.range_partitions;   sys.value_partitions;   sys.dependencies_vw
 

DROP INDEX

   DROP INDEX [ schema_name . ] index_name

Remove an existing table index

Example:

DROP TABLE "my_project"."my_Index";

For details see: Documentation/Manuals/SQLreference/DropStatement
See also: CREATE INDEX   CREATE ORDERED INDEX   CREATE IMPRINTS INDEX

Associated system table: sys.idxs
 

DROP VIEW

   DROP VIEW [ IF EXISTS ] [ schema_name . ] view_name
    [ RESTRICT | CASCADE ]

Remove a view

Example:

DROP VIEW IF EXISTS "my_project"."my_View" CASCADE;

Note: You cannot drop a view if it is referenced (e.g. from another view, a function or procedure or another db object except comments). Use option CASCADE to specify to also drop those referencing objects.
System views can also not be dropped as they are needed by the system.
The "IF EXISTS" option is supported from release Jul2017 (11.27.1) onwards.

For details see: Documentation/Manuals/SQLreference/DropStatement
See also: CREATE VIEW

Associated system table: sys.tables where type = 1
 

DROP AGGREGATE

   DROP AGGREGATE [ IF EXISTS ] [ schema_name . ] aggregate_function_name
    [ ( [ arg1_type [ , arg2_type ] [, ... ] ] ) ]
    [ RESTRICT | CASCADE ]

Remove a specific user-defined aggregation function. If multiple aggregate functions exist with the same name, supply the full signature.

Example:

DROP AGGREGATE testaggr(int, double);

Note: This command is MonetDB specific.
Use option CASCADE to specify to also drop referencing objects.
System aggregates can not be dropped as they are part by the system.
The "IF EXISTS" option is supported from release Mar2018 (11.29.3) onwards.

For details see: Documentation/Manuals/SQLreference/DropStatement
See also: DROP ALL AGGREGATE   CREATE AGGREGATE LANGUAGE   CREATE AGGREGATE EXTERNAL

Associated system table: sys.functions where type = 3
 

DROP ALL AGGREGATE

   DROP ALL AGGREGATE [ schema_name . ] aggregate_function_name
    [ RESTRICT | CASCADE ]

Remove all user-defined aggregation functions which share the same name but have different signatures.

Example:

DROP ALL AGGREGATE testaggr;

Note: This command is MonetDB specific.
Use option CASCADE to specify to also drop referencing objects.

For details see: Documentation/Manuals/SQLreference/DropStatement
See also: DROP AGGREGATE   CREATE AGGREGATE LANGUAGE   CREATE AGGREGATE EXTERNAL

Associated system table: sys.functions where type = 3
 

DROP FUNCTION

   DROP FUNCTION [ IF EXISTS ] [ schema_name . ] function_name
    [ ( [ arg1_type [ , arg2_type ] [, ... ] ] ) ]
    [ RESTRICT | CASCADE ]

Remove a specific user-defined function. If multiple functions exist with the same name, supply the full signature.

Example:

DROP FUNCTION testfun(int, double);

Note: Use option CASCADE to specify to also drop referencing objects.
System functions can not be dropped as they are part by the system.
The "IF EXISTS" option is supported from release Mar2018 (11.29.3) onwards.

For details see: Documentation/Manuals/SQLreference/DropStatement
See also: DROP ALL FUNCTION   CREATE FUNCTION   CREATE FUNCTION LANGUAGE

Associated system table: sys.functions where type in (1,5)
 

DROP ALL FUNCTION

   DROP ALL FUNCTION [ schema_name . ] function_name
    [ RESTRICT | CASCADE ]

Remove all user-defined functions which share the same name but have different signatures.

Example:

DROP ALL FUNCTION testfun;

Note: This command is MonetDB specific.
Use option CASCADE to specify to also drop referencing objects.

For details see: Documentation/Manuals/SQLreference/DropStatement
See also: DROP FUNCTION   CREATE FUNCTION   CREATE FUNCTION LANGUAGE

Associated system table: sys.functions where type in (1,5)
 

DROP FILTER FUNCTION

   DROP FILTER FUNCTION [ IF EXISTS ] [ schema_name . ] function_name
    [ ( [ arg1_type [ , arg2_type ] [, ... ] ] ) ]
    [ RESTRICT | CASCADE ]

Remove a specific user-defined filter function. If multiple filter functions exist with the same name, supply the full signature.

Example:

DROP FILTER FUNCTION my_like(string, varchar(100));

Note: This command is MonetDB specific.
System filter functions can not be dropped as they are part by the system.
Use option CASCADE to specify to also drop referencing objects.
The "IF EXISTS" option is supported from release Mar2018 (11.29.3) onwards.

For details see: Documentation/Manuals/SQLreference/DropStatement
See also: DROP ALL FILTER FUNCTION   CREATE FILTER FUNCTION EXTERNAL

Associated system table: sys.functions where type = 4
 

DROP ALL FILTER FUNCTION

   DROP ALL FILTER FUNCTION [ schema_name . ] function_name
    [ RESTRICT | CASCADE ]

Remove all user-defined filter functions which share the same name but have different signatures.

Example:

DROP ALL FILTER FUNCTION my_like;

Note: This command is MonetDB specific.
Use option CASCADE to specify to also drop referencing objects.

For details see: Documentation/Manuals/SQLreference/DropStatement
See also: DROP FILTER FUNCTION   CREATE FILTER FUNCTION EXTERNAL

Associated system table: sys.functions where type = 4
 

DROP LOADER

   DROP LOADER [ IF EXISTS ] [ schema_name . ] function_name
    [ ( [ arg1_type [ , arg2_type ] [, ... ] ] ) ]
    [ RESTRICT | CASCADE ]

Remove a specific user-defined loader function. If multiple loader functions exist with the same name, supply the full signature.

Note: This command is MonetDB specific. It is introduced in release Dec2016 (11.25.3).
The "IF EXISTS" option is supported from release Mar2018 (11.29.3) onwards.

For details see: blog/monetdbpython-loader-functions
See also: DROP ALL LOADER   CREATE LOADER   CREATE FUNCTION LANGUAGE

Associated system table: sys.functions where type = 7 and language > 2
 

DROP ALL LOADER

   DROP ALL LOADER [ schema_name . ] function_name
    [ RESTRICT | CASCADE ]

Remove all user-defined loader functions which share the same name but have different signatures.

Example:

DROP ALL LOADER my_ods_loader;

Note: This command is MonetDB specific.
Use option CASCADE to specify to also drop referencing objects.

For details see: Documentation/Manuals/SQLreference/DropStatement
See also: DROP LOADER   CREATE LOADER

Associated system table: sys.functions where type = 7 and language > 2
 

DROP PROCEDURE

   DROP PROCEDURE [ IF EXISTS ] [ schema_name . ] procedure_name
    [ ( [ arg1_type [ , arg2_type ] [, ... ] ] ) ]
    [ RESTRICT | CASCADE ]

Remove a specific user-defined procedure. If multiple procedures exist with the same name, supply the full signature.

Example:

DROP PROCEDURE testproc(int, double);

Note: Use option CASCADE to specify to also drop referencing objects.
System procedures can not be dropped as they are part by the system.
The "IF EXISTS" option is supported from release Mar2018 (11.29.3) onwards.

For details see: Documentation/Manuals/SQLreference/DropStatement
See also: DROP ALL PROCEDURE   CREATE PROCEDURE   CREATE PROCEDURE EXTERNAL

Associated system table: sys.functions where type = 2
 

DROP ALL PROCEDURE

   DROP ALL PROCEDURE [ schema_name . ] procedure_name
    [ RESTRICT | CASCADE ]

Remove all user-defined procedures which share the same name but have different signatures.

Example:

DROP ALL PROCEDURE testproc;

Note: This command is MonetDB specific.
Use option CASCADE to specify to also drop referencing objects.

For details see: Documentation/Manuals/SQLreference/DropStatement
See also: DROP PROCEDURE   CREATE PROCEDURE   CREATE PROCEDURE EXTERNAL

Associated system table: sys.functions where type = 2
 

DROP TYPE

   DROP TYPE [ schema_name . ] type_name
    [ RESTRICT | CASCADE ]

Remove a user-defined type

For details see: Documentation/Manuals/SQLreference/Userdefinedtypes
See also: CREATE TYPE EXTERNAL

Associated system table: sys.types
 

DROP TRIGGER

   DROP TRIGGER [ IF EXISTS ] [ schema_name . ] trigger_name

Remove an existing trigger

Note: The "IF EXISTS" option is supported from release Mar2018 (11.29.3) onwards.

For details see: Documentation/Manuals/SQLreference/DropStatement
See also: CREATE TRIGGER

Associated system table: sys.triggers
 

SELECT

   SELECT selection_list
    [ FROM table_view_subselect_function_list [ [ AS ] table_alias ] ]
    [ WINDOW window_definition_list ]
    [ WHERE condition_list ]
    [ GROUP BY grouping_list ]
    [ HAVING group conditon_list ]
    [ { UNION | INTERSECT | EXCEPT } [ ALL | DISTINCT ] [ CORRESPONDING [ BY '(' column_ref_commalist ')' ] ] SELECT_query ]
    [ ORDER BY ordering_list [ ASC | DESC ]  ]
    [ LIMIT nr_of_rows ]
    [ OFFSET row_nr ]
    [ SAMPLE sample_size [ SEED integer ] ]

Execute a database query and return the resultset data rows

Example:

-- query to list all non-system tables:
SELECT tbl.id, table_type_name, sch.name AS "Schema", tbl.name as "Table"
    , (select count(*) from sys.columns where table_id = tbl.id) as nr_of_columns
  FROM sys."tables" as tbl
 INNER JOIN sys.table_types ON tbl.type = table_type_id
 INNER JOIN sys."schemas" as sch ON tbl.schema_id = sch.id
 WHERE NOT tbl.system	-- exclude system tables and views
   AND table_type_name <> 'VIEW'	-- exclude user views
 ORDER by "Schema", "Table";

-- example using window functions
SELECT id, emp_name, dep_name
     , ROW_NUMBER()           OVER (PARTITION BY dep_name ORDER BY id) AS row_number_in_frame
     , NTH_VALUE(emp_name, 2) OVER (PARTITION BY dep_name ORDER BY id) AS second_row_in_frame
     , LEAD(emp_name, 2)      OVER (PARTITION BY dep_name ORDER BY id) AS two_rows_ahead
     , SUM(salary)  OVER (PARTITION BY dep_name ORDER BY salary  ROWS BETWEEN 2 PRECEDING AND CURRENT ROW) AS sum_over_1or2or3_rows
     , SUM(salary)  OVER (PARTITION BY dep_name ORDER BY salary  GROUPS BETWEEN 1 PRECEDING AND 1 FOLLOWING) AS sum_over_groups
     , SUM(salary)  OVER (PARTITION BY dep_name ORDER BY salary  RANGE BETWEEN 100.0 PRECEDING AND 50.0 FOLLOWING) AS sum_over_range
  FROM employee
 ORDER BY dep_name, id;

-- example using (values(tuple),(tuple),...)
select *
 from ( values(0,'cero'),(1,'uno'),(2,'dos'),(3,'tres'),(4,'cuatro'),(5,'cinco'),(6,'seis'),(7,'siete'),(8,'ocho'),(9,'nueve'),(10,'diez') ) as nr_es(nr, nm)

Note: It is also possible to select data without a FROM-clause, such as: SELECT CURRENT_DATE, (1+2) * 3, pi();
The "WINDOW window_definition_list" option and "SEED integer" option are supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/SQLreference/TableExpressions , Documentation/Manuals/SQLreference/WindowFunctions ,
Documentation/Cookbooks/SQLrecipes/Sampling and wiki/Structured_Query_Language/Window_functions

Associated system table: sys.queue and sys.querylog_history
 

SELECT INTO

   SELECT selection_list
    INTO variable_list
    [ FROM table_view_subselect_function_list [ [ AS ] table_alias ] ]
    [ WINDOW window_definition_list ]
    [ WHERE condition_list ]
    [ GROUP BY grouping_list ]
    [ HAVING group conditon_list ]
    [ { UNION | INTERSECT | EXCEPT } [ ALL | DISTINCT ] [ CORRESPONDING [ BY '(' column_ref_commalist ')' ] ] SELECT_query ]
    [ ORDER BY ordering_list [ ASC | DESC ]  ]
    [ LIMIT nr_of_rows ]
    [ OFFSET row_nr ]
    [ SAMPLE sample_size [ SEED integer ] ]

Execute a database query and store the resultset data in the named variables

Note: The "WINDOW window_definition_list" option and "SEED integer" option are supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/SQLreference/TableExpressions
and: Documentation/Cookbooks/SQLrecipes/monitor

Associated system table: sys.queue and sys.querylog_history
 

WITH SELECT

   WITH cte_alias AS SELECT_query
     [, cte_alias AS SELECT_query] [, ... ]
   SELECT selection_list
    [ INTO variable_list ]
    [ FROM cte_alias_table_view_subselect_function_list [ [ AS ] table_alias ] ]
    [ WINDOW window_definition_list ]
    [ WHERE condition_list ]
    [ GROUP BY grouping_list ]
    [ HAVING group conditon_list ]
    [ { UNION | INTERSECT | EXCEPT } [ ALL | DISTINCT ] [ CORRESPONDING [ BY '(' column_ref_commalist ')' ] ] SELECT_query ]
    [ ORDER BY ordering_list [ ASC | DESC ]  ]
    [ LIMIT nr_of_rows ]
    [ OFFSET row_nr ]
    [ SAMPLE sample_size [ SEED integer ] ]

Execute a database query using the results of the queries defined in the WITH clause and return the resultset data rows

Note: Recursive queries are NOT supported yet. This feature request is logged in bugzilla.
The "WINDOW window_definition_list" option and "SEED integer" option are supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/SQLreference/TableExpressions

Associated system table: sys.queue and sys.querylog_history
 

INSERT VALUES

   INSERT INTO [ schema_name . ] table_name
     [ ( column_name [, column_name] [, ... ] ) ]
    VALUES ( values_list )   [, ( values_list ) [, ... ] ]

Add data row(s) to an existing table

Example:

INSERT INTO "mySch"."myTab" (id, name) VALUES (1, 'a'), (2, 'b'), (3, 'c');

Note: It is also possible to add multiple rows of values in one insert statement by adding extra ", (values_list)" parts.

For details see: Documentation/SQLreference/Updates
and: Documentation/Cookbooks/SQLrecipes/LoadingBulkData

Associated system table: sys.querylog_history
 

INSERT SELECT

     [ WITH cte_alias AS SELECT_query [, cte_alias AS SELECT_query] [, ... ] ]
   INSERT INTO [ schema_name . ] table_name
     [ ( column_name [, column_name] [, ... ] ) ]
     SELECT_query

Execute a database query and add the resultset data rows into the existing table

Note: The "WITH cte_alias AS SELECT_query" option is supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/SQLreference/Updates

Associated system table: sys.querylog_history
 

UPDATE

     [ WITH cte_alias AS SELECT_query [, cte_alias AS SELECT_query] [, ... ] ]
   UPDATE [ schema_name . ] table_name [ [ AS ] table_alias ]
     SET column_name = expression_value
        [ , column_name = expression_value ] [, ... ]
     [ FROM cte_alias_table_view_subselect_function_list [ [ AS ] table_alias ] ]
     [ WHERE condition_expression ]

Change values of specific column(s) and row(s) of an existing table. If the optional WHERE clause is specified only the rows which match the condition(s) are changed. Table alias option is available since Apr2019 release.

Example:

UPDATE employee SET salary = salary * 1.07, bonus = 1200 WHERE id = 12345;

Note: The "WITH cte_alias AS SELECT_query" option is supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/SQLreference/Updates

Associated system table: sys.querylog_history
 

MERGE

     [ WITH cte_alias AS SELECT_query [, cte_alias AS SELECT_query] [, ... ] ]
   MERGE INTO [ schema_name . ] table_name [ [ AS ] target_alias ]
    USING table_or_cte_ref [ [ AS ] source_alias ]
       ON merge_search_condition
     WHEN merge_action
     [ WHEN merge_action ]

  merge_action:

     NOT MATCHED THEN INSERT [ (column1 [, column2 ...] ) ] [ { VALUES (value1 [, value2 ...] ) | DEFAULT VALUES } ]
   | MATCHED THEN UPDATE SET column1 = value1 [, column2 = value2 ] [, ... ]
   | MATCHED THEN DELETE

The MERGE INTO command is used to make changes in one table (target) based on values matched from another (source). It can be used to combine insert and update or insert and delete operations into one command, instead of multiple INSERT and UPDATE/DELETE commands. You must specify a merge search condition which determines whether a record will be inserted or updated/deleted.

Example:

MERGE INTO ProductTarget T
     USING ProductSource S
        ON S.ProductID = T.ProductID
      WHEN MATCHED     THEN UPDATE SET Name = S.Name, ProductNumber = S.ProductNumber, Color = S.Color
      WHEN NOT MATCHED THEN INSERT (ProductID, Name, ProductNumber, Color)
                            VALUES (S.ProductID, S.Name, S.ProductNumber, S.Color);

Note: This SQL:2003 command is supported from release Apr2019 (11.33.3) onwards.
Limitations: Multiple WHEN MATCHED clauses or multiply WHEN NOT MATCHED clauses in a merge command are not allowed.
Also WHEN MATCHED AND condition THEN ... syntax is not supported.

For details see: merge_statements_now_supported   and   en.wikipedia.org/wiki/Merge_(SQL)

Associated system table: sys.querylog_history
 

DELETE

     [ WITH cte_alias AS SELECT_query [, cte_alias AS SELECT_query] [, ... ] ]
   DELETE FROM [ schema_name . ] table_name [ [ AS ] table_alias ]
     [ WHERE condition_expression ]

Remove data row(s) from an existing table. If the optional WHERE clause is specified only the rows which match the condition(s) are removed. Table alias option is available since Apr2019 release.

Example:

DELETE FROM web.log WHERE "timestamp" <= '2016-12-23 23:59:59';

Note: The "WITH cte_alias AS SELECT_query" option is supported from release Apr2019 (11.33.3) onwards.

For details see: Documentation/SQLreference/Updates
See also: TRUNCATE

Associated system table: sys.querylog_history
 

TRUNCATE

   TRUNCATE [ TABLE ] [ schema_name . ] table_name
    [ CONTINUE IDENTITY | RESTART IDENTITY ]
    [ RESTRICT | CASCADE ]

Remove all data row(s) from an existing table quickly

Example:

TRUNCATE TABLE mysch.imp_article CONTINUE IDENTITY CASCADE;

Note: The option RESTART IDENTITY can be used to reset the sequence start value of an identity column for new data inserts after the truncate. Default behavior is to CONTINUE IDENTITY sequence numbering.
The CASCADE option instructs to truncate referencing table(s) also if the referencing table(s) have foreign key references to this table. The default behavior is RESTRICT.
This command is supported from release Mar2018 (11.29.3) onwards.
Note: it is possible to use TRUNCATE statements in a transaction and thus to rollback the effects of a truncate.

For details see: Documentation/SQLreference/Updates
See also: DELETE

Associated system table: sys.querylog_history
 

CALL

   CALL [ schema_name . ] procedure_name ( [ arg1_val [ , arg2_val ] [, ... ] ] )

Execute a stored procedure

Example:

CALL "sys"."settimeout"(3000);

Note: The current user must have EXECUTE privilege for the called procedure.

For details see: Documentation/SQLreference/Flowofcontrol
See also: CREATE PROCEDURE   GRANT PRVILEGES

Associated system table: sys.querylog_calls
 

PREPARE

   PREPARE SQL_DML_statement-with-optional-question_mark-parameter-markers

Compiles a SQL statement into its execution plan. This is useful for statements which need to be executed many times but with different values, such as an INSERT command.

Example:

PREPARE INSERT INTO towns (name, country, citizens) VALUES (?, ?, ?);
EXECUTE 5('Amsterdam', 'NL', 856124);
EXECUTE 5('Berlin', 'DE', 3715930);

PREPARE SELECT id, name, schema_id FROM tables WHERE name like ?;
EXECUTE 7('%type%');
EXECUTE 7('%id%');

Note: You can prepare any DML command, such as SELECT, INSERT, UPDATE, DELETE, MERGE, CALL, etc.
The prepared statement will be given a system identifier number which can be used in the EXECUTE command.
Instead of keyword PREPARE you may also use PREP.

For details see: Documentation/Manuals/SQLreference/PrepareExec
and: Documentation/Cookbooks/SQLrecipes/LoadingBulkData
See also: EXECUTE

Associated system table: sys.querylog_catalog
 

EXECUTE

   EXECUTE prepared-SQL_nr ( [ parm1_val [ , parm2_val ] [, ... ] ] )

Execute the prepared statement with parameter values

Example:

EXECUTE 3('c-data');
EXECUTE 5('Amsterdam', 'NL', 856124);

Note: Instead of keyword EXECUTE you may also use EXEC.
WARNING: when an error occurs in the user session, all the user prepared statement(s) are lost!

For details see: Documentation/Manuals/SQLreference/PrepareExec
See also: PREPARE

Associated system table: sys.querylog_calls
 

SET

   SET variable_name = new_value

Change the value of a declared variable

For details see: Documentation/Manuals/SQLreference/Variables
See also: DECLARE   SELECT

Associated system table: sys.var()
 

SET CURRENT_TIMEZONE

   SET CURRENT_TIMEZONE = minutes

Change the current timezone offset from GMT

Example:

SELECT CURRENT_TIMEZONE;
SET CURRENT_TIMEZONE = 3600;

For details see: Documentation/Manuals/SQLreference/Variables
See also: CURRENT_TIMEZONE

Associated system table: sys.var_values
 

SET TIME ZONE

   SET TIME ZONE { LOCAL | INTERVAL 'hh:mi' HOUR TO MINUTE }

Change the current timezone

Example:

SELECT CURRENT_TIMEZONE;
SET TIME ZONE LOCAL;
SELECT CURRENT_TIMEZONE;
SET TIME ZONE INTERVAL '+02:00' HOUR TO MINUTE;
SET TIME ZONE INTERVAL '3' HOUR;
SET TIME ZONE INTERVAL '240' MINUTE;
SET TIME ZONE INTERVAL '-3600' SECOND;

For details see: Documentation/Manuals/SQLreference/Variables    Documentation/SQLreference/Temporal
 

CREATE ROLE

   CREATE ROLE role_name
    [ WITH ADMIN grantor ]

Define a new role. You can use the role to grant privileges and next grant a role (or multiple roles) to specific users. This eases the maintenance of specifying privileges to new users or changing privileges for a set of users which have the same role.

Example:

CREATE ROLE controller;

For details see: Documentation/SQLreference/Roles
and: Documentation/SQLreference/Permissions
See also: DROP ROLE   GRANT PRIVILEGES   GRANT ROLE   SET ROLE

Associated system table: sys.roles and sys.auths
 

CREATE USER

   CREATE USER user_login_name
    WITH [ ENCRYPTED | UNENCRYPTED ] PASSWORD pass_phrase
    NAME full_name
    SCHEMA schema_name

Define a new database user account. Only monetdb administrator may create new database users.

Example:

CREATE USER "Donald" WITH UNENCRYPTED PASSWORD 'Secret' NAME 'Donald Duck' SCHEMA "app1";

For details see: Documentation/SQLreference/Users
and: Documentation/SQLreference/Permissions
See also: ALTER USER RENAME   ALTER USER SET PASSWORD   GRANT PRIVILEGES   DROP USER

Associated system table: sys.users and sys.auths where name in (select name from sys.users)
 

ALTER USER RENAME

   ALTER USER user_login_name
    RENAME TO new_user_login_name

Change the user id name of a database user account

For details see: Documentation/SQLreference/Users
See also: ALTER USER SET PASSWORD   ALTER USER SET SCHEMA   GRANT PRIVILEGES

Associated system table: sys.users
 

ALTER USER SET PASSWORD

   ALTER USER SET [ ENCRYPTED | UNENCRYPTED ] PASSWORD new_pass_phrase
    USING OLD PASSWORD old_pass_phrase

Change the password of the current user account

Example to change the default password of system user monetdb (when connected as monetdb):

SELECT CURRENT_USER;
ALTER USER SET PASSWORD 'make it private' USING OLD PASSWORD 'monetdb';

For details see: Documentation/SQLreference/Users
See also: ALTER USER SET SCHEMA   ALTER USER RENAME   GRANT PRIVILEGES

Associated system table: sys.users
 

ALTER USER SET SCHEMA

   ALTER USER user_login_name
    [ WITH [ ENCRYPTED | UNENCRYPTED ] PASSWORD pass_phrase ]
    SET SCHEMA schema_name

Change the default schema of a database user account

For details see: Documentation/SQLreference/Users
See also: ALTER USER SET PASSWORD   ALTER USER RENAME   GRANT PRIVILEGES

Associated system table: sys.users and sys.auths
 

DROP ROLE

   DROP ROLE role_name
    [ WITH ADMIN grantor ]

Remove an existing role

Example:

DROP ROLE controller;

Note: Associated role privileges and granted user_role records will also be removed from sys.privileges and sys.user_role

For details see: Documentation/SQLreference/Roles
and: Documentation/Manuals/SQLreference/DropStatement
See also: CREATE ROLE   GRANT ROLE   SET ROLE

Associated system table: sys.roles and sys.user_role

DROP USER

   DROP USER user_login_name

Remove an existing database login account

Note: Associated user privileges and granted user_role records will also be removed from sys.privileges and sys.user_role

For details see: Documentation/SQLreference/Users
and: Documentation/Manuals/SQLreference/DropStatement
See also: CREATE USER   ALTER USER RENAME

Associated system table: sys.users and sys.auths where name in (select name from sys.users)

GRANT PRIVILEGES

   GRANT { SELECT | INSERT | UPDATE | DELETE | TRUNCATE | REFERENCES | ALL [ PRIVILEGES ] | EXECUTE | COPY INTO | COPY FROM }
    [, ... ]
    ON [ TABLE | FUNCTION | AGGREGATE ] object_name
    TO { user_login_name | role_name | PUBLIC } [, ... ]
    [ WITH GRANT OPTION ]

Add privileges on a table or function (incl. procedure, filter function, loader) or aggregate function for a user or role or all users (PUBLIC).
For tables the privileges can be: SELECT, INSERT, UPDATE, DELETE, TRUNCATE, REFERENCES, ALL, COPY INTO, COPY FROM. ALL implies privileges: SELECT, INSERT, UPDATE, DELETE, TRUNCATE and REFERENCES. Note: privilege REFERENCES is currently implemented as a synonym for SELECT.
For aggregates, functions and procedures the privilege can be: EXECUTE or ALL.
The privileges COPY INTO, COPY FROM enable execution of those COPY INTO, COPY FROM commands to users other than the monetdb system user.

Example:

GRANT ALL ON TABLE web_event TO PUBLIC;
GRANT SELECT, INSERT ON TABLE "web_log" TO jan, piet, controller WITH GRANT OPTION;
GRANT EXECUTE ON FUNCTION refresh_func TO PUBLIC;
GRANT COPY INTO, COPY FROM ON TABLE new_facts TO mrdata;

Note: The current user must have permission to GRANT privileges.
The COPY INTO and COPY FROM privileges are MonetDB specific SQL. They are introduced in release Jun2016 (11.23.3).

For details see: Documentation/SQLreference/Permissions
See also: REVOKE PRIVILEGES   GRANT ROLE   CREATE USER

Associated system table: sys.privileges
 

REVOKE PRIVILEGES

   REVOKE [ GRANT OPTION FOR ]
    { SELECT | INSERT | UPDATE | DELETE | TRUNCATE | REFERENCES | ALL [ PRIVILEGES ] | EXECUTE | COPY INTO | COPY FROM }
    [, ... ]
    ON [ TABLE | FUNCTION | AGGREGATE ] object_name
    FROM { user_login_name | role_name | PUBLIC } [, ... ]

Remove privileges on a table or function (incl. procedure, filter function, loader) or aggregate function from a user or role or PUBLIC

Example:

REVOKE INSERT, DELETE, TRUNCATE ON TABLE web_event FROM jan;
REVOKE EXECUTE ON FUNCTION refresh_func FROM piet;

Note: The current user must have permission to GRANT privileges.
The COPY INTO and COPY FROM privileges are MonetDB specific SQL. They are introduced in release Jun2016 (11.23.3).

For details see: Documentation/SQLreference/Permissions
See also: GRANT PRIVILEGES   REVOKE ROLE   DROP USER

Associated system table: sys.privileges
 

GRANT ROLE

   GRANT role_name [, ... ] TO user_login_name [, ... ]
    [ WITH ADMIN OPTION ]
    [ WITH ADMIN grantor ]

Add one or more role privileges to one or more users

Example:

GRANT controller TO jan WITH ADMIN OPTION;

For details see: Documentation/SQLreference/Permissions
and: Documentation/SQLreference/Roles
See also: REVOKE ROLE   GRANT PRIVILEGES   CREATE ROLE

Associated system table: sys.user_role
 

REVOKE ROLE

   REVOKE [ ADMIN OPTION FOR ] role_name [, ... ] FROM user_login_name [, ... ]

Remove one or more role privileges from one or more users

Example:

REVOKE controller FROM jan;

For details see: Documentation/SQLreference/Permissions
and: Documentation/SQLreference/Roles
See also: GRANT ROLE   DROP ROLE

Associated system table: sys.user_role
 

SET SCHEMA

   SET SCHEMA schema_name

Change the current schema

Example:

SELECT CURRENT_SCHEMA;
SET SCHEMA profiler;
SELECT CURRENT_SCHEMA;

For details see: Documentation/SQLreference/Schema
and: Documentation/Manuals/SQLreference/Variables
See also: CURRENT_SCHEMA

Associated system table: sys.var_values
 

SET ROLE

   SET ROLE role_name

Change the current role

Example:

SELECT CURRENT_ROLE;
SET ROLE controller;
SELECT CURRENT_ROLE;

For details see: Documentation/SQLreference/Roles
and: Documentation/Manuals/SQLreference/Variables
See also: CURRENT_ROLE   GRANT ROLE

Associated system table: sys.var_values
 

SET USER

   SET USER = user_login_name

Change the current user to the specified user

Example:

SELECT USER;
SET USER = jan;
SELECT USER;

Note: USER is a synonym for CURRENT_USER

For details see: Documentation/SQLreference/Users
and: Documentation/Manuals/SQLreference/Variables
See also: USER   GRANT PRIVILEGES

Associated system table: sys.var_values
 

SET CURRENT_USER

   SET CURRENT_USER = user_login_name

Change the current user to the specified user

Example:

SELECT CURRENT_USER;
SET CURRENT_USER = piet;
SELECT CURRENT_USER;

For details see: Documentation/SQLreference/Users
and: Documentation/Manuals/SQLreference/Variables
See also: CURRENT_USER

Associated system table: sys.var_values
 

SET SESSION_USER

   SET SESSION_USER = user_login_name

Change the session user to the specified user

Example:

SET SESSION_USER = isabel;
SELECT SESSION_USER;

For details see: Documentation/SQLreference/Users
and: Documentation/Manuals/SQLreference/Variables
See also: SESSION_USER

Associated system table: sys.var_values
 

SET SESSION AUTHORIZATION

   SET SESSION AUTHORIZATION role_name

Change the session authorization to the specified role

For details see: Documentation/SQLreference/Roles
and: Documentation/Manuals/SQLreference/Variables
 

COPY INTO FROM

   COPY [ int_val [ OFFSET int_val ] RECORDS ]
    INTO [ schema_name . ] table_name
         [ '(' column_name [ ',' column_name [ ... ] ] ')' ]
    FROM file_name [ ',' file_name [ ... ] ]
     [ '(' ident [ STRING ] [ ',' ident [ STRING ] [ ... ] ] ')' ]
     [ ON { CLIENT | SERVER } ]
     [ [ USING ] DELIMITERS  field_separator
       [ ',' record_separator [ ',' string_quote ] ] ]
     [ NULL [ AS ] null_string ]
     [ LOCKED ]
     [ BEST EFFORT ]
     [ NO CONSTRAINT ]
     [ FWF '(' pos [ ',' pos [ ... ] ] ')' ]

Read data from an external csv/txt file and bulk insert it into an existing table

Note: This command is MonetDB specific.
By default the file(s) must be accessible from the server where mserver5 process is running. From release Apr2019 (11.33.3) onwards it is possible to read the file(s) from the mclient program by specifying ON CLIENT. Default behavior is ON SERVER. The ON CLIENT option is not supported in other programs or APIs (ODBC, JDBC, python DBI, etc.).
It can also read directly from a compressed file if the file name has extension: .bz2 or .gz or .xz or .lz4. Extension .zip is not recognised, use extension .gz instead.
The default field_separator is the pipe '|' character.
The default record_separator is the newline '\n' character.
The default string_quote character is a double quote: ".
IMPORTANT: All character string data in the csv/txt file must be stored in UTF-8 character encoding. Other character encodings are not supported.

For details see: Documentation/Cookbooks/SQLrecipes/CSV_bulk_loads
and: Documentation/Cookbooks/SQLrecipes/LoadingBulkData

Associated system table: sys.querylog_history
 

COPY INTO FROM STDIN

   COPY [ int_val [ OFFSET int_val ] RECORDS ]
    INTO [ schema_name . ] table_name
         [ '(' column_name [ ',' column_name [ ... ] ] ')' ]
    FROM STDIN
     [ '(' ident [ STRING ] [ ',' ident [ STRING ] [ ... ] ] ')' ]
     [ [ USING ] DELIMITERS  field_separator
       [ ',' record_separator [ ',' string_quote ] ] ]
     [ NULL [ AS ] null_string ]
     [ LOCKED ]
     [ BEST EFFORT ]
     [ NO CONSTRAINT ]

Read data from console (stdin) and bulk insert it into an existing table

Note: This command is MonetDB specific.
The default field_separator is the pipe '|' character.
The default record_separator is the newline '\n' character.
The default string_quote character is a double quote: ".

For details see: Documentation/Cookbooks/SQLrecipes/CSV_bulk_loads
and: Documentation/Cookbooks/SQLrecipes/LoadingBulkData
and: SQLcopyinto.java example program

Associated system table: sys.querylog_history
 

COPY BINARY INTO FROM

   COPY BINARY
    INTO [ schema_name . ] table_name
         [ '(' column_name [ ',' column_name [ ... ] ] ')' ]
    FROM file_name [ ',' file_name [ ... ] ]
     [ ON { CLIENT | SERVER } ]
     [ NO CONSTRAINT ]

Attach column data from an external binary file(s) to very quickly make it available in the RDBMS

Note: This command is MonetDB specific.
By default the files must be accessible from the server where mserver5 process is running. From release Apr2019 (11.33.3) onwards it is possible to read the file from the mclient program by specifying ON CLIENT. Default behavior is ON SERVER. The ON CLIENT option is not supported in other programs or APIs (ODBC, JDBC, python DBI, etc.).
IMPORTANT: not all datatypes are supported. Also a NULL value must be represented as a special value, which is different per datatype.

For details see: Documentation/Cookbooks/SQLrecipes/BinaryBulkLoad

Associated system table: sys.querylog_history
 

COPY LOADER INTO FROM

   COPY LOADER INTO [ schema_name . ] table_name
    FROM function_name ( [ arg1_val [ , arg2_val ] [, ... ] ] )

Read data via a loader function and bulk insert it into an existing table

Example:

COPY LOADER INTO tbl FROM web_log_loader();

Note: This command is MonetDB specific. It is introduced in release Dec2016 (11.25.3).
For a loader function defined in language PYTHON the Python script interpreter software must be available on the deployment server.

For details see: blog/monetdbpython-loader-functions
See also: CREATE LOADER   CREATE TABLE FROM LOADER

Associated system table: sys.querylog_history
 

COPY INTO

   COPY SELECT_query
    INTO file_name
    [ ON { CLIENT | SERVER } ]
    [ [ USING ] DELIMITERS field_separator
      [ ',' record_separator [ ',' string_quote ] ] ]
    [ NULL AS null_string ]

Write query result data in bulk to an external csv/txt file

Note: This command is MonetDB specific.
By default the file path must be accessible from the server where mserver5 process is running. From release Apr2019 (11.33.3) onwards it is now possible to write the file locally by the mclient program by specifying ON CLIENT. Default behavior is ON SERVER. The ON CLIENT option is not supported in other programs or APIs (ODBC, JDBC, python DBI, etc.).
The default field_separator is the pipe '|' character.
The default record_separator is the newline '\n' character.
The default string_quote character is a double quote: ".
Character data will be exported in UTF-8 encoding. Other encodings are not supported.
The file can also be written directly as compressed file if the file name has extension: .bz2 or .gz or .xz or .lz4. Extension .zip is not recognised, use extension .gz instead.

For details see: Documentation/Cookbooks/SQLrecipes/ExportingBulkData
and: Documentation/Manuals/SQLreference/CopyInto

Associated system table: sys.querylog_history
 

COPY INTO STDOUT

   COPY SELECT_query
    INTO STDOUT
    [ [ USING ] DELIMITERS field_separator
      [ ',' record_separator [ ',' string_quote ] ] ]
    [ NULL AS null_string ]

Write query result data to console (stdout)

Note: This command is MonetDB specific.
The default field_separator is the pipe '|' character.
The default record_separator is the newline '\n' character.
The default string_quote character is a double quote: ".

For details see: Documentation/Cookbooks/SQLrecipes/ExportingBulkData
and: Documentation/Manuals/SQLreference/CopyInto

Associated system table: sys.querylog_history
 

SET TRANSACTION

   SET TRANSACTION
    [ READ ONLY | READ WRITE ]  [ ISOLATION LEVEL { READ UNCOMMITTED | READ COMMITTED | REPEATABLE READ | SERIALIZABLE } ]

Start a transaction (turn auto-commit off) with optionally a transaction mode or isolation level

Example:

SET TRANSACTION;
ROLLBACK;
SET TRANSACTION READ ONLY;
ROLLBACK;
SET TRANSACTION READ WRITE;
ROLLBACK;
SET TRANSACTION ISOLATION LEVEL READ UNCOMMITTED;
ROLLBACK;
SET TRANSACTION ISOLATION LEVEL READ COMMITTED;
ROLLBACK;
SET TRANSACTION ISOLATION LEVEL REPEATABLE READ;
ROLLBACK;
SET TRANSACTION ISOLATION LEVEL SERIALIZABLE;
ROLLBACK;

SET LOCAL TRANSACTION

   SET LOCAL TRANSACTION
    [ READ ONLY | READ WRITE ]  [ ISOLATION LEVEL { READ UNCOMMITTED | READ COMMITTED | REPEATABLE READ | SERIALIZABLE } ]

Start a local transaction (turn auto-commit off) with optionally a transaction mode or isolation level

 

START TRANSACTION

   START TRANSACTION
    [ READ ONLY | READ WRITE ] [ ISOLATION LEVEL { READ UNCOMMITTED | READ COMMITTED | REPEATABLE READ | SERIALIZABLE } ]

Start a user controlled transaction

Note: A transaction can also include data definition (DDL) commands such as CREATE, ALTER, DROP.

For details see: Documentation/Manuals/SQLreference/Transactions
and: blog/monetdb-sql-transaction-management-scheme
See also: COMMIT   ROLLBACK
 

COMMIT

   COMMIT [ WORK ]

Make all changes done since the start of the transaction persistent

For details see: Documentation/Manuals/SQLreference/Transactions
and: blog/monetdb-sql-transaction-management-scheme
See also: ROLLBACK   START TRANSACTION
 

ROLLBACK

   ROLLBACK [ WORK ]

Undo all changes done since the start of the transaction

For details see: Documentation/Manuals/SQLreference/Transactions
and: blog/monetdb-sql-transaction-management-scheme
See also: COMMIT   START TRANSACTION
 

SAVEPOINT

   SAVEPOINT savepoint_id_name

Add a savepoint within a transaction

Example:

START TRANSACTION;
SAVEPOINT sp1;

Note: You need to start a transaction first before you can use savepoints

See also: START TRANSACTION   RELEASE SAVEPOINT   ROLLBACK TO SAVEPOINT
 

ROLLBACK TO SAVEPOINT

   ROLLBACK TO SAVEPOINT savepoint_id_name

Undo all changes done since the start of the savepoint

Example:

ROLLBACK TO SAVEPOINT sp1;

See also: SAVEPOINT   RELEASE SAVEPOINT
 

RELEASE SAVEPOINT

   RELEASE SAVEPOINT savepoint_id_name

Add all changes done since the start of the savepoint to the current transaction

Example:

RELEASE SAVEPOINT sp1;

See also: SAVEPOINT   ROLLBACK TO SAVEPOINT
 

SET OPTIMIZER

   SET optimizer = optimizer_pipeline_name

Change the execution optimizer pipeline

Example:

SELECT optimizer;
SELECT * FROM sys.optimizers;
SET optimizer = 'sequential_pipe';
SELECT optimizer;
SET optimizer = 'default_pipe';
SELECT optimizer;
SET optimizer = 'ceci_nest_pas_une_pipe';
SELECT optimizer;

For details see: Documentation/Cookbooks/SQLrecipes/OptimizerPipelines
and: Documentation/Manuals/MonetDB/Optimizers
See also: PLAN   EXPLAIN

Associated system table: sys.optimizers and sys.environment where name = 'sql_optimizer';
 

EXPLAIN

   EXPLAIN The SQL command

Show the MAL execution plan for the SQL command.

For details see: Documentation/Manuals/SQLreference/Explain    Documentation/Manuals/MonetDB/MALreference
and: Documentation/Cookbooks/SQLrecipes/OptimizerPipelines
See also: PLAN
 

PLAN

   PLAN The SQL command

Show the relational execution plan for the SQL command after it has been optimized by the SQL optimzer.

Note: This command is MonetDB specific.

For details see: Documentation/Manuals/SQLreference/PlanSQL
and: Documentation/Cookbooks/SQLrecipes/OptimizerPipelines
See also: EXPLAIN
 

TRACE

   TRACE The SQL command

Executes the SQL command and return a table with the execution ticks of the MAL instructions.

For details see: Documentation/Manuals/SQLreference/Trace

Associated system table: sys.tracelog()
 

DEBUG

   DEBUG The SQL command

Executes the SQL command in DEBUG mode

For details see: Documentation/Manuals/SQLreference/Debug
 

ANALYZE

   ANALYZE schema_name [ . table_name [ ( column_name [, column_name] [, ... ] ) ] ]
    [ SAMPLE size ]
    [ MINMAX ]

Gather statistics on column(s) of table(s) in a specific schema and store or update them in the system table sys.statistics for analysis

Example:

-- derive the statistics for table sys.statistics
ANALYZE sys.statistics;
SELECT * FROM sys.statistics;

-- update statistics of specific columns of sys.statistics
ANALYZE sys.statistics ("sample", count, "unique", minval, maxval) SAMPLE 1000;

SELECT * FROM sys.statistics
 WHERE column_id IN (SELECT id FROM sys.columns
 WHERE table_id  IN (SELECT id FROM sys.tables
 WHERE schema_id IN (SELECT id FROM sys.schemas
 WHERE name = 'sys') AND name = 'statistics'));

-- remove all the statistics
DELETE FROM sys.statistics;

Note: The statistics information can be used by the query optimizer to choose the optimal execution plan. Rerun the ANALYZE command after many table manipulations (insert/update/delete).
For large tables this command can take some time. Specify a SAMPLE size to reduce the amount of data values to analyze. Add MINMAX to exclude the expensive unique count operation.

For details see: Documentation/Cookbooks/SQLrecipes/statistics

Associated system table: sys.statistics
 

USER

   USER

Example:

SELECT USER;

Note: USER is a synonym for CURRENT_USER

See also: SET USER

Associated system table: sys.var_values
 

CURRENT_USER

   CURRENT_USER

Example:

SELECT CURRENT_USER;

See also: SET CURRENT_USER

Associated system table: sys.var_values
 

SESSION USER

   SESSION_USER

Example:

SELECT SESSION_USER;

See also: SET SESSION_USER

Associated system table: sys.var_values
 

CURRENT_ROLE

   CURRENT_ROLE

Example:

SELECT CURRENT_ROLE;

See also: SET ROLE

Associated system table: sys.var_values
 

CURRENT_SCHEMA

   CURRENT_SCHEMA

Example:

SELECT CURRENT_SCHEMA;

See also: SET SCHEMA

Associated system table: sys.var_values
 

CURRENT_TIMEZONE

   CURRENT_TIMEZONE

Example:

SELECT CURRENT_TIMEZONE;

For details see: Documentation/Cookbooks/SQLrecipes/DateTimeFunctionality
See also: SET CURRENT_TIMEZONE

Associated system table: sys.var_values
 

NOW

   NOW

Example:

SELECT NOW;

For details see: Documentation/Cookbooks/SQLrecipes/DateTimeFunctionality
 

CURRENT_DATE

   CURRENT_DATE

Example:

SELECT CURRENT_DATE;

For details see: Documentation/Cookbooks/SQLrecipes/DateTimeFunctionality
 

CURRENT_TIME

   CURRENT_TIME

Example:

SELECT CURRENT_TIME;

For details see: Documentation/Cookbooks/SQLrecipes/DateTimeFunctionality
 

CURRENT_TIMESTAMP

   CURRENT_TIMESTAMP

Example:

SELECT CURRENT_TIMESTAMP;

For details see: Documentation/Cookbooks/SQLrecipes/DateTimeFunctionality
 

LOCALTIME

   LOCALTIME

Example:

SELECT LOCALTIME;

For details see: Documentation/Cookbooks/SQLrecipes/DateTimeFunctionality
 

LOCALTIMESTAMP

   LOCALTIMESTAMP

Example:

SELECT LOCALTIMESTAMP;

For details see: Documentation/Cookbooks/SQLrecipes/DateTimeFunctionality
 

NEXT VALUE FOR

   NEXT VALUE FOR sequence_name

Example:

CREATE SEQUENCE "my_test_seq" as integer START WITH 2;
CREATE TABLE test (t int DEFAULT NEXT VALUE FOR "my_test_seq", v varchar);
INSERT INTO test(v) VALUES ('ab');
SELECT * FROM test;

For details see: Documentation/Manuals/SQLreference/SerialTypes
 

Comments

Comments can be added to query scripts for documentation purposes. MonetDB/SQL supports two forms of comments. Any text beginning with '- -' and up to the end of line is ignored. Furthermore, C-style comments ( /* this is ignored */ ) can be injected in a SQL query where a language token is expected. A language token can be a keyword, identifier, a literal constant or a special character. Tokens should be  separated by 'white space' (tab, space, newlines) unless there juxtaposition does not create an ambiguous interpretation. Comments are considered equivalent to white spaces.

Identifiers and Keywords

SQL comes with a large collection of keywords, i.e. names reserved by the committee to designate language constructs. Keywords in MonetDB/SQL are case in-sensitive and adhere to the general convention for identifier denotation in programming languages.

Users can overrule the interpretation of an identifier as a keyword by encapsulation with double quotes, e.g. select denotes a keyword, while "select" denotes a user defined name. This scheme also permits inclusion of white space in the names. However, it is general advisable to limit the number of the escaped keywords.

Names are used to designate database objects. In that role, they are case in-sensitive. The terminal identifier classes distinguished are denoted with the suffix _name.

Literal Constants

Numeric constants follow the convention of most programming languages. A numeric constant that contains neither a decimal point or exponential is considered of type integer. The underlying machine determines whether it will be interpreted as a 32- or 64-bit value, although most likely it will be that latter case these days.

String constants in SQL are embraced with single quotes like: 'string data'. If your data contains single quote's, each single quote need to be doubled, so 'a single '' quote in a string' represents string: a single ' quote in a string. We support Unicode character string literals (from Apr2019 release 11.33.3) using: U&'...' including UESCAPE. For the string literals, you can have U&'...' '...' '...' UESCAPE '...' where the escape must be as single character and the other '...' strings are also Unicode character string literals. For now, these latter strings also undergo C-style backslash interpretation.

We also support PostgreSQL-like E'...' strings (from Apr2019 release 11.33.3). The strings can contain C-style backslash escapes.

Often strings can be cast to other types using the CAST( expr AS type) expression, provided the typed coercion routine is available. For example:

CAST ( '129' AS integer)
CAST ( '0.3' AS double)

illustrates conversion of a color value in hexadecimal string notation into an integer, and a value into a double precision floating point number.

Temporal Constants

The DATE, TIME and TIMESTAMP data types come with a straightforward string literal conversion structure, e.g.:

SELECT DATE '2014-02-03', TIME '15:45:56', TIMESTAMP '2014-02-03 15:45:56';

produces a single properly typed row result. It is equivalent to:

SELECT CAST('2014-02-03' AS DATE), CAST ('15:45:56' AS TIME), CAST ('2014-02-03 15:45:56' AS TIMESTAMP);

Special Characters

String literals (currently both the '...' and E'...' styles) may contain the C-style escape characters: '\n' for new lines, '\t' for tab, '\r' for return, and '\\' for backslash. The conventions '\ddd` where d a digit stands for a number denoted in octal.

Operator Precedences

Most operators in SQL have the same precedence and are left-associative. Parenthesis can be used to disambiguate the precedence order. The operator definitions can not be overloaded, nor can you define new operators.

Value Expressions

The language is built around value- and table- expressions. Value expressions encompass denotation of literal constants, type casts, column references, operator and function invocation, and scalar returning subqueries.

The column references take the form [correlation_name '.'] column_name. The correlation_name is either a table name or an alias introduced in a from clause. The correlation name may be omitted if the column name uniquely identifies a column within the scope of current query.

The table references take the form [schema_name '.'] table_or_view_name [AS alias_name]. Table expressions produce a relational table. It is internally referenced through a correlation name, which supports attribute selection using the '.' denotation.

SQL prescribes a large collection of built-in types, most of which are directly supported. The MonetDB database kernel can also be extended with user defined types to better support application domains.

Some domain specific data types (e.g. GeoSpatial) are not part of the MonetDB core, but are provided by extensions. For more information check the Spatial data types article.

The MonetDB/SQL supports the full range of object definitions.

MonetDB provides the following groups of built-in SQL functions and operators.

SQL comes with  imperative programming concepts, such as variable declarations, compound statements, while-loops, and conditional-statements. The are grouped into procedures, functions, and triggers, which strongly differ in their execution paradigm and semantics.

MonetDB/SQL supports a multi-statement transaction scheme marked by START TRANSACTION and closed with either COMMIT or ROLLBACK. The session variable AUTOCOMMIT can be set to true (default) if each SQL statement should be considered an independent transaction.

In the AUTOCOMMIT mode, you can use START TRANSACTION and COMMIT/ROLLBACK to indicate transactions containing multiple SQL statements.  In this case, AUTOCOMMIT is automatically disabled by a START TRANSACTION, and reenabled by a COMMIT or ROLLBACK.

If AUTOCOMMIT mode is OFF, the START TRANSACTION is implicit, and you should only use COMMIT/ROLLBACK.

WARNING. The transaction management scheme is based on optimistic concurrency control. It provides each transaction with a consistent view on the database, but updates are collected in an addendum processed on transaction commit. If at commit time it can be assured that the data prepared for update affects tables has not changed in the mean time, the results are merged. Otherwise the transaction is aborted.
This optimistic concurrency scheme is particularly useful for query dominant environments. It negatively affects long running transactions which concurrently are affected by updates on their underlying tables. The same holds for applications that tries to perform concurrent updates from multiple threads in a single application. They should be internally serialized by the application to avoid unexpected transaction aborts.

WARNING. Optimistic concurrency control may be confusing for those who built online-transaction applications, because the granularity of the concurrency control scheme will show higher then expected transaction failures. There is not a locking schema to avoid this. Applications may have to resort to serial execution.

WARNING. The tuples being deleted are only marked as such. They do not reduce the table size. You will even experience after many updates that your queries will run slower, because each query first has to establish a consistent private view on the database by inspection of the delete/update lists. It calls for a vacuum cleaning algorithm in the background, which is not yet available.

The SQL statement modifiers provide system specific information over e.g. relational query plan, MAL execution plan, performance, and access to the debugger.

MonetDB comes with JDBC, ODBC, PHP, Perl, Ruby and Python interface libraries. The JDBC, PHP, Perl, Ruby, Python and Node.js interfaces are native implementations and do not require installation of the MonetDB client/server code. The ODBC driver and setup library are available as a separate installer. The Mapi library is the lowest level C-interface to interacting with the server.

We rely on external documentation for the basics of the language bindings provided: PHP, Perl, Python, and the drivers JDBC and ODBC. Examples provided are used to illustrate their behavior in the context of MonetDB only.

Notes: The programming interface is based on a client-server architecture, where the client program connects to a server using a TCP/IP connection to exchange commands and receives answers. The underlying protocol uses plain UTF-8 data for ease of use and debugging. This leads to publicly visible information exchanged over a network, which may be undesirable. Therefore, a private and secure channel can be set up with the Secure Socket Layer functionality.

In addition, a deeper integration of MonetDB with R is available on the website of the MonetDB.R connector. If you are looking for information about the embedded R support in MonetDB, follow check this blog post.

More information about using MonetDB on Node.js you can find here.