`CREATE` statement

The CREATE statement can be used to add a record to the database. If the record already exists, the statement will give an error.

Note
This statement can not be used to create graph relationships. For that, use the RELATE statement.

Statement syntax

SurrealQL Syntax
CREATE [ ONLY ] @targets
	[ CONTENT @value
	  | SET @field = @value ...
	]
	[ RETURN NONE | RETURN BEFORE | RETURN AFTER | RETURN DIFF | RETURN @statement_param, ... | RETURN VALUE @statement_param ]
	[ TIMEOUT @duration ]
	[ PARALLEL ]
;

Creating a Table Record

CREATE can be used with just a table name, in which case its ID will be generated randomly.

-- Create a new record
CREATE person;

Response
[
    {
        "id": "person:2vvgzt6m24s952yiy7x8"
    }
]

To specify a specific ID for a table instead, use : followed by a value.

CREATE person:one;

Response
[
	{
		id: person:one
	}
]

The table name and ID together form the full record ID which can be used to query the created data or by using the SELECT statement. See the record ID page to learn more about what counts as a valid record identifier.

The default random ID can be generated in different ways (such as a ULID) using the built-in ID generation functions.

It is also possible to specify the ID of the record you want to create using a string or any of the supported formats for record IDs.

-- Use the type::thing() function to provide a record's table and id separately
CREATE type::thing("person", "one");

Adding Record Data

When creating a record, you can specify the record data using the SET clause, or the CONTENT clause. The SET clause is used to specify record data one field at a time, while the CONTENT clause is used to specify record data using a SurrealQL object. The CONTENT clause is useful when the record data is already in the form of a SurrealQL or JSON object.

-- Create a new record with a text id
CREATE person:tobie SET
    name = 'Tobie',
    company = 'SurrealDB',
    skills = ['Rust', 'Go', 'JavaScript'];

The above will create a new record with the ID person:tobie and the specified data.

Response
[
	{
		"id": "person:tobie",
		"name": "Tobie",
		"company": "SurrealDB",
		"skills": ["Rust", "Go", "JavaScript"]
	}
]

Specifing record data using the CONTENT keyword:

-- Create a new record with a numeric id
CREATE person:100 CONTENT {
	name: 'Tobie',
	company: 'SurrealDB',
	skills: ['Rust', 'Go', 'JavaScript'],
};

Options and clauses

Creating multiple records

Multiple records or even multiple record types can be created by separating table names by commas.

-- Note: meta::tb(id) returns just the table name portion of a record ID
CREATE townsperson, cat, dog SET
    created_at = time::now(),
    name = "Just a " + meta::tb(id);

Response
[
    {
        "created_at": "2024-03-19T03:12:05.079Z",
        "id": "townsperson:p37ha2lngckp3v8tvf2j",
        "name": "Just a townsperson"
    },
    {
        "created_at": "2024-03-19T03:12:05.080Z",
        "id": "cat:p1pwbjaq96nhhnuohjtc",
        "name": "Just a cat"
    },
    {
        "created_at": "2024-03-19T03:12:05.080Z",
        "id": "dog:01vcxgdpuctdk354hzkp",
        "name": "Just a dog"
    }
]

The | | syntax is another way to create multiple records in a single execution. This syntax can be used in two ways.

One is by including a table name, a : (a colon), and then a number. This will create a quantity of records equal to the number after the table name. The records created will have random IDs.

-- Creates three townperson records with a random ID
CREATE |townsperson:3|;

Response
[
	{
		id: townsperson:hzkt0piy3f72xo5dl2jf
	},
	{
		id: townsperson:k0mujrohm8qe2txz5pnz
	},
	{
		id: townsperson:pwumqelrsi1qt0jmihwh
	}
]

The other method is by using the .. range syntax after the : instead of a single number. This will create records with specific IDs that span from the lower to the upper range.

CREATE |townsperson:1..3|;

Response
[
	{
		id: townsperson:1
	},
	{
		id: townsperson:2
	},
	{
		id: townsperson:3
	}
]

All of these methods can be combined to create multiple records at the same time.

CREATE dog, |cat:2|, |townsperson:1..3| SET
    created_at = time::now(),
    name = "Just a " + meta::tb(id);

Response
[
	{
		created_at: '2024-08-13T04:14:44.135Z',
		id: dog:u3fzmqvg3yq9mo3o6z2s,
		name: 'Just a dog'
	},
	{
		created_at: '2024-08-13T04:14:44.137Z',
		id: cat:n6x3caiiazucslfs7rpm,
		name: 'Just a cat'
	},
	{
		created_at: '2024-08-13T04:14:44.137Z',
		id: cat:rnvhxgjhsbea5u58s0wu,
		name: 'Just a cat'
	},
	{
		created_at: '2024-08-13T04:14:44.137Z',
		id: townsperson:1,
		name: 'Just a townsperson'
	},
	{
		created_at: '2024-08-13T04:14:44.137Z',
		id: townsperson:2,
		name: 'Just a townsperson'
	},
	{
		created_at: '2024-08-13T04:14:44.137Z',
		id: townsperson:3,
		name: 'Just a townsperson'
	}
]

ONLY

When creating a single record, the ONLY clause can be used to return the record object on its own instead of inside an array.

-- Returns an array with a single record inside
CREATE person:tobie SET
    name = 'Tobie',
    company = 'SurrealDB',
    skills = ['Rust', 'Go', 'JavaScript'];

-- Returns just a single record
CREATE ONLY person:tobie SET
    name = 'Tobie',
    company = 'SurrealDB',
    skills = ['Rust', 'Go', 'JavaScript'];

Response
-------- Query --------

[
	{
		company: 'SurrealDB',
		id: person:tobie,
		name: 'Tobie',
		skills: [
			'Rust',
			'Go',
			'JavaScript'
		]
	}
]

-------- Query --------

{
	company: 'SurrealDB',
	id: person:tobieagain,
	name: 'Tobie',
	skills: [
		'Rust',
		'Go',
		'JavaScript'
	]
}

Return Values

By default, the create statement returns the record once it has been created. To change what is returned, we can use the RETURN clause, specifying either NONE, BEFORE, AFTER, DIFF, or a comma-separated list of specific fields to return.

RETURN NONE can be useful to avoid excess output:

-- Create 10000 records but don't show any of them
CREATE |person:10000| SET age = 46, username = "john-smith" RETURN NONE;

RETURN DIFF returns the changeset diff:

CREATE person SET age = 46, username = "john-smith" RETURN DIFF;

Response
[
	[
		{
			op: 'replace',
			path: '/',
			value: {
				age: 46,
				id: person:h84x4k5kh2m6cjf1vvza,
				username: 'john-smith'
			}
		}
	]
]

RETURN BEFORE inside a CREATE statement is essentially a synonym for RETURN NONE, while RETURN AFTER is the default behaviour for create.

-- Will always return NONE
CREATE person SET age = 46, username = "john-smith" RETURN BEFORE;

-- Return the record after creation
CREATE person SET age = 46, username = "john-smith" RETURN AFTER;

You can also return specific fields from a created record, the value of a single field using VALUE, as well as ad-hoc fields to modify the output as needed.

CREATE person
    SET age = 46,
    username = "john-smith",
    interests = ['skiing', 'music']
RETURN
    age,
    interests,
    age + 1 AS age_next_year;

CREATE |person:5|
    SET age = 20
RETURN VALUE age;

Response
-------- Query --------

[
	{
		age: 46,
		age_next_year: 47,
		interests: [
			'skiing',
			'music'
		]
	}
]

-------- Query --------

[
	20,
	20,
	20,
	20,
	20
]

Timeout

The TIMEOUT clause can be used to specify the maximum time the statement should take to execute. This is useful when you want more control such as controlling compute costs or making sure queries succeed or fail within tight latency boundaries to not have a big query queue forming.

The value for TIMEOUT is specified in seconds or milliseconds.

-- Query attempting to create half a million `person` records
CREATE |person:500000| SET age = 46, username = "john-smith" TIMEOUT 500ms;

Parallel

The PARALLEL keyword can be used to specify that the statement should be processed concurrently, rather than sequentially. Similar to the TIMEOUT clause this is useful for more control over how your queries should behave, if that is needed.

CREATE person:26, CREATE person:27 PARALLEL;

`VERSION`

Available since: v2.0.0

If you are using SurrealKV as the storage engine, when creating a record you can specify a version for each record. This is useful for time-travel queries. You can query a specific version of a record by using the VERSION clause.

The VERSION clause is always followed by a datetime and when the specified timestamp does not exist, an empty array is returned.

Note
The VERSION clause is currently in alpha and is subject to change. We do not recommend this for production.

-- Create a record for user:john at 8:00AM
CREATE user:john SET name = 'John' VERSION d'2024-08-19T08:00:00Z';
[[{ id: user:john, name: 'John' }]]

-- Return the record for user:john at 8:00AM
SELECT * FROM user:john VERSION d'2024-08-19T08:00:00Z';
[[{ id: user:john, name: 'John' }]]

-- Create a record for user:john at 8:01AM
CREATE user:john SET name = 'John-1' VERSION d'2024-08-19T08:01:00Z';
[[{ id: user:john, name: 'John-1' }]]

-- Return the record for user:john at 8:01AM
SELECT * FROM user:john VERSION d'2024-08-19T08:01:00Z';
[[{ id: user:john, name: 'John-1' }]]

-- Return an empty array because the record at the datetime does not exist
SELECT * FROM user:john VERSION d'2024-08-19T07:00:00Z';
[[]]

Another example of how VERSION works with CREATE is by creating records at different times and then querying for them at a specific point in time.

CREATE |user:10| VERSION d"2020-09-09";

[[{ id: user:rtbjoqv1xe9wnxjx5aro }, { id: user:tkik878q8uoddvuucu0a }, { id: user:rcnywgogvlipv3tb8qut }, { id: user:30ynx82x52ff77dxzv1i }, { id: user:59mxi0xosi3im5ccbx8l }, { id: user:nolu7yreqs4e5m7255oa }, { id: user:u384ycj1d2esi3yrasli }, { id: user:n4xnrq98ookevhmdd7d2 }, { id: user:5j5ujfu4dokcpdk51qa8 }, { id: user:jiqmlvrgafeorr50nvn9 }]]

CREATE |user:10| VERSION d"2020-09-10";
[[{ id: user:ze98ow4bzdcndzc5nlqj }, { id: user:gjqu2uh3wnp1cpjg1unt }, { id: user:17bxpjl4ptbxv9k2ghmt }, { id: user:fmqqeajf52neg4c7oaoq }, { id: user:bfn45ewsg86auvekeuz0 }, { id: user:834yq1tyatwopb4726mj }, { id: user:veehoua4cu65ff4wc8pf }, { id: user:y3az4pizc0ddpruixw6g }, { id: user:xrn6eqrtyqgg8cgpm9zp }, { id: user:s06acf74rsnvhvim3ys5 }]]

RETURN count(SELECT * FROM user VERSION d"2020-09-09"); -- returns 10
[10]

RETURN count(SELECT * FROM user); -- returns 21
[21]

Implicit statement behaviour

While a number of definitions need to be in place for a CREATE statement to happen, SurrealDB will handle them automatically by default. This behaviour is best seen by starting a new database.

While a connection to SurrealDB via Surrealist or the surreal sql command can include a defined namespace and database, the namespace and database names do not exist upon creation. At this point, they are only held inside the pre-defined $session parameter. This can be seen through the INFO statements, which will show no definitions at all inside a new database.

INFO FOR ROOT;
INFO FOR NS;
INFO FOR DB;
RETURN $session;

Response
-------- Query --------

{
	namespaces: {},
	users: {}
}

-------- Query --------

{
	databases: {},
	tokens: {},
	users: {}
}

-------- Query --------

{
	analyzers: {},
	functions: {},
	models: {},
	params: {},
	scopes: {},
	tables: {},
	tokens: {},
	users: {}
}

-------- Query --------

{
	db: 'sandbox',
	exp: NONE,
	id: NONE,
	ip: NONE,
	ns: 'sandbox',
	or: NONE,
	sc: NONE,
	sd: NONE,
	tk: NONE
}

This is to allow the chance to define them manually, such as by including a comment.

DEFINE DATABASE my_database COMMENT "Some important info that I prefer to add manually";

However, once the first record is created or inserted, SurrealDB will access the session data to execute a number of definition statements for the namespace, database, and then add a definition for the desired table name in order to allow the operation to proceed.

-- Three DEFINE statements will happen to allow this operation
CREATE person;

INFO FOR ROOT;
INFO FOR NS;
INFO FOR DB;

Response
-------- Query --------

{
	namespaces: {
		sandbox: 'DEFINE NAMESPACE sandbox'
	},
	users: {}
}

-------- Query --------

{
	databases: {
		sandbox: 'DEFINE DATABASE sandbox'
	},
	tokens: {},
	users: {}
}

-------- Query 7 --------

{
	analyzers: {},
	functions: {},
	models: {},
	params: {},
	scopes: {},
	tables: {
		person: 'DEFINE TABLE person TYPE ANY SCHEMALESS PERMISSIONS NONE'
	},
	tokens: {},
	users: {}
}

To disallow this behaviour, you can pass the --strict flag when starting the database. In strict mode, everything must first be explicitly defined before it can be used.

ns/db> CREATE person;
["The namespace 'ns' does not exist"]

ns/db> DEFINE NAMESPACE ns;
[NONE]

ns/db> CREATE person;
["The database 'db' does not exist"]

ns/db> DEFINE DATABASE db;
[NONE]

ns/db> CREATE person;
["The table 'person' does not exist"]

ns/db> DEFINE TABLE person;
[NONE]

ns/db> CREATE person;
[[{ id: person:c76lfw6n4yb1z2dj9xaj }]]

Learn more

To learn more about SurrealDB, check out the following resources:

Edit this page on GitHub

CONTINUE DEFINE

SurrealQL

CREATE statement