MongoDB

MongoDB Sink Connector

Key features​

• exactly-once

• cdc

Tips

1.If you want to use CDC-written features, recommend enable the upsert-enable configuration.

Description​

The MongoDB Connector provides the ability to read and write data from and to MongoDB. This document describes how to set up the MongoDB connector to run data writers against MongoDB.

Data Type Mapping​

The following table lists the field data type mapping from MongoDB BSON type to Nexus data type.

Nexus Data Type	MongoDB BSON Type
STRING	ObjectId
STRING	String
BOOLEAN	Boolean
BINARY	Binary
INTEGER	Int32
TINYINT	Int32
SMALLINT	Int32
BIGINT	Int64
DOUBLE	Double
FLOAT	Double
DECIMAL	Decimal128
Date	Date
Timestamp	Timestamp[Date]
ROW	Object
ARRAY	Array

Tips

1.When using Nexus to write Date and Timestamp types to MongoDB, both will produce a Date data type in MongoDB, but the precision will be different. The data generated by the Nexus Date type has second-level precision, while the data generated by the Nexus Timestamp type has millisecond-level precision. 2.When using the DECIMAL type in Nexus, be aware that the maximum range cannot exceed 34 digits, which means you should use decimal(34, 18).

Sink Options​

Name	Type	Required	Default	Description
uri	String	Yes	-	The MongoDB standard connection uri. eg. mongodb://user:password@hosts:27017/database?readPreference=secondary&slaveOk=true.
database	String	Yes	-	The name of MongoDB database to read or write.
collection	String	Yes	-	The name of MongoDB collection to read or write.
schema	String	Yes	-	MongoDB's BSON and Nexus data structure mapping.
buffer-flush.max-rows	String	No	1000	Specifies the maximum number of buffered rows per batch request.
buffer-flush.interval	String	No	30000	Specifies the maximum interval of buffered rows per batch request, the unit is millisecond.
retry.max	String	No	3	Specifies the max number of retry if writing records to database failed.
retry.interval	Duration	No	1000	Specifies the retry time interval if writing records to database failed, the unit is millisecond.
upsert-enable	Boolean	No	false	Whether to write documents via upsert mode.
primary-key	List	No	-	The primary keys for upsert/update. Keys are in ["id","name",...] format for properties.
transaction	Boolean	No	false	Whether to use transactions in MongoSink (requires MongoDB 4.2+).
common-options		No	-	Sink plugin common parameters, please refer to Sink Common Optionsmon Options for details

Tips​

1.The data flushing logic of the MongoDB Sink Connector is jointly controlled by three parameters: buffer-flush.max-rows, buffer-flush.interval, and checkpoint.interval. Data flushing will be triggered if any of these conditions are met. 2.Compatible with the historical parameter upsert-key. If upsert-key is set, please do not set primary-key.

How to Create a MongoDB Data Synchronization Jobs​

The following example demonstrates how to create a data synchronization job that writes randomly generated data to a MongoDB database:

# Set the basic configuration of the task to be performed
env {
  parallelism = 1
  job.mode = "BATCH"
  checkpoint.interval  = 1000
}

source {
  FakeSource {
      row.num = 2
      bigint.min = 0
      bigint.max = 10000000
      split.num = 1
      split.read-interval = 300
      schema {
        fields {
          c_bigint = bigint
        }
      }
    }
}

sink {
  MongoDB{
    uri = mongodb://user:[email protected]:27017
    database = "test"
    collection = "test"
    schema = {
      fields {
        _id = string
        c_bigint = bigint
      }
    }
  }
}

Parameter Interpretation​

MongoDB Database Connection URI Examples​

Unauthenticated single node connection:

mongodb://127.0.0.0:27017/mydb

Replica set connection:

mongodb://127.0.0.0:27017/mydb?replicaSet=xxx

Authenticated replica set connection:

mongodb://admin:[email protected]:27017/mydb?replicaSet=xxx&authSource=admin

Multi-node replica set connection:

mongodb://127.0.0..1:27017,127.0.0..2:27017,127.0.0.3:27017/mydb?replicaSet=xxx

Sharded cluster connection:

mongodb://127.0.0.0:27017/mydb

Multiple mongos connections:

mongodb://192.168.0.1:27017,192.168.0.2:27017,192.168.0.3:27017/mydb

Note: The username and password in the URI must be URL-encoded before being concatenated into the connection string.

Buffer Flush​

sink {
  MongoDB {
    uri = "mongodb://user:[email protected]:27017"
    database = "test_db"
    collection = "users"
    buffer-flush.max-rows = 2000
    buffer-flush.interval = 1000
    schema = {
      fields {
        _id = string
        id = bigint
        status = string
      }
    }
  }
}

Why is Not Recommended to Use Transactions for Operation?​

Although MongoDB has fully supported multi-document transactions since version 4.2, it doesn't mean that everyone should use them recklessly. Transactions are equivalent to locks, node coordination, additional overhead, and performance impact. Instead, the principle for using transactions should be: avoid using them if possible. The necessity for using transactions can be greatly avoided by designing systems rationally.

Idempotent Writes​

By specifying a clear primary key and using the upsert method, exactly-once write semantics can be achieved.

If primary-key and upsert-enable is defined in the configuration, the MongoDB sink will use upsert semantics instead of regular INSERT statements. We combine the primary keys declared in upsert-key as the MongoDB reserved primary key and use upsert mode for writing to ensure idempotent writes. In the event of a failure, Nexus jobs will recover from the last successful checkpoint and reprocess, which may result in duplicate message processing during recovery. It is highly recommended to use upsert mode, as it helps to avoid violating database primary key constraints and generating duplicate data if records need to be reprocessed.

sink {
  MongoDB {
    uri = "mongodb://user:[email protected]:27017"
    database = "test_db"
    collection = "users"
    upsert-enable = true
    primary-key = ["name","status"]
    schema = {
      fields {
        _id = string
        name = string
        status = string
      }
    }
  }
}