Docs Home → Develop Applications → MongoDB Manual
create![](/zh-cn/docs/v4.4/assets/link.svg)
On this page
Definition![](/zh-cn/docs/v4.4/assets/link.svg)
create
Explicitly creates a collection or view.
Note
The view created by this command does not refer to materialized views. For discussion of on-demand materialized views, see
$merge
instead.create
has the following form:Note
Starting in MongoDB 4.2
MongoDB removes the MMAPv1 storage engine and the MMAPv1 specific option
flags
forcreate
.{ create: <collection or view name>, capped: <true|false>, autoIndexId: <true|false>, size: <max_size>, max: <max_documents>, storageEngine: <document>, validator: <document>, validationLevel: <string>, validationAction: <string>, indexOptionDefaults: <document>, viewOn: <source>, pipeline: <pipeline>, collation: <document>, writeConcern: <document>, comment: <any> } create
has the following fields:FieldTypeDescriptioncreate
stringThe name of the new collection or view. See Naming Restrictions.capped
booleanOptional. To create a capped collection, specifytrue
. If you specifytrue
, you must also set a maximum size in thesize
field.autoIndexId
booleanOptional. Specify
false
to disable the automatic creation of an index on the_id
field.Important
Starting in MongoDB 4.0, you cannot set the option
autoIndexId
tofalse
when creating collections in databases other than thelocal
database.Deprecated since version 3.2.
size
integerOptional. Specify a maximum size in bytes for a capped collection. Once a capped collection reaches its maximum size, MongoDB removes the older documents to make space for the new documents. Thesize
field is required for capped collections and ignored for other collections.max
integerOptional. The maximum number of documents allowed in the capped collection. Thesize
limit takes precedence over this limit. If a capped collection reaches thesize
limit before it reaches the maximum number of documents, MongoDB removes old documents. If you prefer to use themax
limit, ensure that thesize
limit, which is required for a capped collection, is sufficient to contain the maximum number of documents.storageEngine
documentOptional. Available for the WiredTiger storage engine only.
Allows users to specify configuration to the storage engine on a per-collection basis when creating a collection. The value of the
storageEngine
option should take the following form:{ <storage-engine-name>: <options> } Storage engine configuration specified when creating collections are validated and logged to the oplog during replication to support replica sets with members that use different storage engines.
validator
documentOptional. Allows users to specify validation rules or expressions for the collection. For more information, see Schema Validation.
New in version 3.2.
The
validator
option takes a document that specifies the validation rules or expressions. You can specify the expressions using the same operators as the query operators with the exception of$near
,$nearSphere
,$text
, and$where
.Note
Validation occurs during updates and inserts. Existing documents do not undergo validation checks until modification.
You cannot specify a validator for collections in the
admin
,local
, andconfig
databases.You cannot specify a validator for
system.*
collections.
validationLevel
stringOptional. Determines how strictly MongoDB applies the validation rules to existing documents during an update.
New in version 3.2.
validationLevel
Description"off"
No validation for inserts or updates."strict"
Default Apply validation rules to all inserts and all updates."moderate"
Apply validation rules to inserts and to updates on existing valid documents. Do not apply rules to updates on existing invalid documents.validationAction
stringOptional. Determines whether to
error
on invalid documents or justwarn
about the violations but allow invalid documents to be inserted.New in version 3.2.
Important
Validation of documents only applies to those documents as determined by the
validationLevel
.validationAction
Description"error"
Default Documents must pass validation before the write occurs. Otherwise, the write operation fails."warn"
Documents do not have to pass validation. If the document fails validation, the write operation logs the validation failure.indexOptionDefaults
documentOptional. Allows users to specify a default configuration for indexes when creating a collection.
The
indexOptionDefaults
option accepts astorageEngine
document, which should take the following form:{ <storage-engine-name>: <options> } Storage engine configuration specified when creating indexes are validated and logged to the oplog during replication to support replica sets with members that use different storage engines.
New in version 3.2.
viewOn
stringThe name of the source collection or view from which to create the view. The name is not the full namespace of the collection or view; i.e. does not include the database name and implies the same database as the view to create. You must create views in the same database as the source collection.
See also
db.createView()
.New in version 3.4.
pipeline
arrayAn array that consists of the aggregation pipeline stage(s).
create
creates the view by applying the specifiedpipeline
to theviewOn
collection or view.The view definition
pipeline
cannot include the$out
or the$merge
stage. If the view definition includes nested pipeline (e.g. the view definition includes$lookup
or$facet
stage), this restriction applies to the nested pipelines as well.The view definition is public; i.e.
db.getCollectionInfos()
andexplain
operations on the view will include the pipeline that defines the view. As such, avoid referring directly to sensitive fields and values in view definitions.See also
db.createView()
.New in version 3.4.
collation
Specifies the default collation for the collection or the view.
Collation allows users to specify language-specific rules for string comparison, such as rules for lettercase and accent marks.
The collation option has the following syntax:
collation: { locale: <string>, caseLevel: <boolean>, caseFirst: <string>, strength: <int>, numericOrdering: <boolean>, alternate: <string>, maxVariable: <string>, backwards: <boolean> } When specifying collation, the
locale
field is mandatory; all other collation fields are optional. For descriptions of the fields, see Collation Document.If you specify a collation at the collection level:
Indexes on that collection will be created with that collation unless the index creation operation explicitly specify a different collation.
Operations on that collection use the collection's default collation unless they explicitly specify a different collation.
You cannot specify multiple collations for an operation. For example, you cannot specify different collations per field, or if performing a find with a sort, you cannot use one collation for the find and another for the sort.
If no collation is specified for the collection or for the operations, MongoDB uses the simple binary comparison used in prior versions for string comparisons.
For a view, if no collation is specified, the view's default collation is the "simple" binary comparison collator. For a view on a collection, the view does not inherit the collection's collation settings. For a view on another view, the to be created view must specify the same collation settings.
After you create the collection or the view, you cannot update its default collation.
For an example that specifies the default collation during the creation of a collection, see Specify Collation.
New in version 3.4.
writeConcern
documentOptional. A document that expresses the write concern for the operation. Omit to use the default write concern.
When issued on a sharded cluster,
mongos
converts the write concern of thecreate
command and its helperdb.createCollection()
to"majority"
.comment
anyOptional. A user-provided comment to attach to this command. Once set, this comment appears alongside records of this command in the following locations:
mongod log messages, in the
attr.command.cursor.comment
field.Database profiler output, in the
command.comment
field.currentOp
output, in thecommand.comment
field.
A comment can be any valid BSON type (string, integer, object, array, etc).
New in version 4.4.
The
db.createCollection()
method and thedb.createView()
method wrap thecreate
command.
Behavior![](/zh-cn/docs/v4.4/assets/link.svg)
Resource Locking![](/zh-cn/docs/v4.4/assets/link.svg)
Changed in version 4.2.
create
obtains an exclusive lock on the
specified collection or view for the duration of the operation. All
subsequent operations on the collection must wait until
create
releases the lock. create
typically holds
this lock for a short time.
Creating a view requires obtaining an additional exclusive lock
on the system.views
collection in the database. This lock blocks
creation or modification of views in the database until the command
completes.
Prior to MongoDB 4.2, create
obtained an exclusive lock
on the parent database, blocking all operations on the database and
all its collections until the operation completed.
Transactions![](/zh-cn/docs/v4.4/assets/link.svg)
Changed in version 4.4.
You can create collections and indexes inside a distributed transaction if the transaction is not a cross-shard write transaction.
To use create
in a transaction, the transaction must use read
concern "local"
. If you specify a read concern level
other than "local"
, the transaction fails.
Tip
Access Control![](/zh-cn/docs/v4.4/assets/link.svg)
If the deployment enforces
authentication/authorization,
create
requires the following privileges:
Task | Required Privileges |
---|---|
Create a non-capped collection |
|
Create a capped collection |
|
Create a view |
However, if the user has the |
A user with the readWrite
built in role on the database
has the required privileges to run the listed operations. Either
create a user with the required role
or grant the role to an existing user.
Examples![](/zh-cn/docs/v4.4/assets/link.svg)
Create a Capped Collection![](/zh-cn/docs/v4.4/assets/link.svg)
To create a capped collection limited to 64 kilobytes, issue the command in the following form:
db.runCommand( { create: "collection", capped: true, size: 64 * 1024 } )
Create a View![](/zh-cn/docs/v4.4/assets/link.svg)
Note
The view created by this command does not refer to materialized
views. For discussion of on-demand materialized views, see
$merge
instead.
Changed in version 4.2.
The view definition pipeline
cannot
include the $out
or the $merge
stage. If the view definition includes
nested pipeline (e.g. the view definition includes
$lookup
or $facet
stage), this
restriction applies to the nested pipelines
as well.
To create a view using the create
command, use the following syntax:
db.runCommand( { create: <view>, viewOn: <source>, pipeline: <pipeline> } )
or if specifying a collation:
db.runCommand( { create: <view>, viewOn: <source>, pipeline: <pipeline>, collation: <collation> } )
For example, given a collection survey
with the following documents:
{ _id: 1, empNumber: "abc123", feedback: { management: 3, environment: 3 }, department: "A" } { _id: 2, empNumber: "xyz987", feedback: { management: 2, environment: 3 }, department: "B" } { _id: 3, empNumber: "ijk555", feedback: { management: 3, environment: 4 }, department: "A" }
The following operation creates a managementRatings
view with the
_id
, feedback.management
, and department
fields:
db.runCommand ( { create: "managementFeedback", viewOn: "survey", pipeline: [ { $project: { "management": "$feedback.management", department: 1 } } ] } )
Important
The view definition is public; i.e. db.getCollectionInfos()
and explain
operations on the view will include the pipeline that
defines the view. As such, avoid referring directly to sensitive fields
and values in view definitions.
Tip
See also:
Specify Collation![](/zh-cn/docs/v4.4/assets/link.svg)
You can specify collation at the collection or view level. For example, the following operation creates a collection, specifying a collation for the collection (See Collation Document for descriptions of the collation fields):
db.runCommand ( { create: "myColl", collation: { locale: "fr" } });
This collation will be used by indexes and operations that support
collation unless they explicitly specify a different collation. For
example, insert the following documents into myColl
:
{ _id: 1, category: "café" } { _id: 2, category: "cafe" } { _id: 3, category: "cafE" }
The following operation uses the collection's collation:
db.myColl.find().sort( { category: 1 } )
The operation returns documents in the following order:
{ "_id" : 2, "category" : "cafe" } { "_id" : 3, "category" : "cafE" } { "_id" : 1, "category" : "café" }
The same operation on a collection that uses simple binary collation (i.e. no specific collation set) returns documents in the following order:
{ "_id" : 3, "category" : "cafE" } { "_id" : 2, "category" : "cafe" } { "_id" : 1, "category" : "café" }
Specify Storage Engine Options![](/zh-cn/docs/v4.4/assets/link.svg)
You can specify collection-specific storage engine configuration
options when you create a collection with
db.createCollection()
. Consider the following operation:
db.runCommand( { create: "users", storageEngine: { wiredTiger: { configString: "<option>=<setting>" } } } )
This operation creates a new collection named users
with a
specific configuration string that MongoDB will pass to the
wiredTiger
storage engine. See the WiredTiger documentation of
collection level options
for specific wiredTiger
options.