Sink data from RisingWave to MongoDB

This guide describes how to sink data from RisingWave to MongoDB. MongoDB is a document database designed for ease of application development and scaling. For more information, see MongoDB.

Syntax

CREATE SINK [ IF NOT EXISTS ] sink_name
[FROM sink_from | AS select_query]
WITH (
   connector='mongodb',
   connector_parameter = 'value', ...
);

Parameters

Parameter Name	Description
mongodb.url	The URL of MongoDB.
type	Defines the type of the sink. Options include append-only or upsert.
collection.name	The collection name where data should be written to or read from. For sinks, the format is db_name.collection_name. Data can also be written to dynamic collections, see collection.name.field below for more information.
collection.name.field	Optional. The dynamic collection name where data should be sunk to. If specified, the field value will be used as the collection name. The collection name format is the same as collection.name. If the field value is null or an empty string, then the collection.name will be used as a fallback destination.
collection.name.field.drop	Optional. Controls whether the field value of collection.name.field should be dropped when sinking. Set this option to true to avoid the duplicate values of collection.name.field being written to the result collection.

Data type mapping

MongoDB Type	RisingWave Type
Boolean	BOOLEAN
32-bit integer	SMALLINT
32-bit integer	INTEGER
64-bit integer	BIGINT
Double	REAL
Double	DOUBLE
Decimal128	DECIMAL
String	DATE
String	VARCHAR
String	TIME
Date	TIMESTAMP WITHOUT TIME ZONE
Date	TIMESTAMP WITH TIME ZONE
String	INTERVAL
Object	STRUCT
Array	ARRAY
Binary data	BYTEA
Object	JSONB
64-bit integer	SERIAL

Examples

Below are some use cases for your reference.

Sink data with append-only

To create a sink with the append-only type:

CREATE sink t1_sink FROM t1
WITH (
    connector='mongodb',
    type = 'append-only',
    mongodb.url = 'mongodb://mongodb:27017/?replicaSet=rs0',
    collection.name = 'demo.t1'
);

In append-only mode, MongoDB will automatically generate an _id field for each record, typically with a value of the ObjectId type. This is necessary because _id is the primary key in MongoDB.

Sink data with upsert

To create a sink with the upsert type for a table with a single key:

single key

CREATE sink t2_sink FROM t2
WITH (
    connector='mongodb',
    type = 'upsert',
    mongodb.url = 'mongodb://mongodb:27017/?replicaSet=rs0',
    collection.name = 'demo.t2',
    primary_key='id'
);

Assuming the schema of t2 is:

name	type	pk
id	int	✔
value	text

Given the record:

id	value
1	’example of record’

The record written to MongoDB will be:

{ "_id": 1, "id": 1, "value": "example of record" }

No redundant id field will exist if the primary key of t2 is _id.

compound key

CREATE TABLE t3(
    a int,
    b int,
    value text,
    primary key (a, b)
);

insert into t3 values(1, 2, 'abc');

CREATE sink t3_sink FROM t3
WITH (
    connector='mongodb',
    type = 'upsert',
    mongodb.url = 'mongodb://mongodb:27017/?replicaSet=rs0',
    collection.name = 'demo.t3',
    primary_key='a,b'
);

The record written to MongoDB will be:

{ "_id": { "a": 1, "b": 2 }, "a": 1, "b": 2, "value": "abc" }

Dynamic collection name

Dynamic collection names are useful in certain scenarios. For example, a multi-tenant application may store its data using sharding, where tenant_id is included as a prefix in the collection name, such as sharding_2024_01.tenant1_order. This approach offers more flexibility and enables efficient data organization and retrieval based on specific tenant requirements.

To use a dynamic collection name:

CREATE sink t2_sink FROM t2
WITH (
    connector='mongodb',
    type = 'upsert',
    mongodb.url = 'mongodb://mongodb:27017/?replicaSet=rs0',
    collection.name = 'demo.t2',
    collection.name.field = 'collection_name',
    collection.name.field.drop = 'true',
    primary_key='_id'
);

• collection.name: Serve as a fallback collection name if the value of collection.name.field is empty or null. In this case, it defaults to demo.t2.
• collection.name.field: Specify the field used for the collection name. This field must be of type varchar.
• collection.name.field.drop: When set to true, it avoids duplicate values of collection.name.field in the result collection.