MongoUser Guide

Overview

Mongo is a pure Fantom driver for MongoDB.

Mongo driver features:

Developed against the MongoDB Stable API v1 from MongoDB 5.2+ (compatible with MongoDB 3.6+)
Support for multi-document, multi-collection transactions
Support for retryable read and write operations
Support for Replica Set connection URLs
Pooled Connection Manager for multi-threaded use and automatic topology scanning and connection fail over
Pluggable authentication with a default SCRAM-SHA-1 over SASL implementation
zlib wire compression

Mongo API features:

Standard CRUD commands: insert(), find(), update(), replace(), and delete()
Special commands: count(), aggregate(), findAndUpdate(), findAndDelete()
Index support commands: create(), ensure() and drop()
Large query cursor support
Text indexes and text searching
Simplified Query DSL syntax
ALIEN-AID: See Morphia for a complete Fantom to MongoDB object mapping library!

Quick Start

Start up an instance of MongoDB:

C:\> mongod

MongoDB starting
db version v5.2.0
waiting for connections on port 27017

Create a text file called Example.fan

using afBson::BsonIO
using afMongo

class Example {

    Void main() {
        mongoClient := MongoClient(`mongodb://localhost:27017`)
        collection  := mongoClient.db("friends").collection("birds")

        documentIn  := [
            "_id"   : ObjectId(),
            "name"  : "Emma",
            "score" : 9
        ]
        collection.insert(documentIn)

        emma        := collection.find.toList.first

        echo("Emma:")
        echo(BsonIO().print(emma, 20))

        mongoClient.shutdown
    }
}

Run Example.fan as a Fantom script from the command line:

C:\> fan Example.fan

      Fantom-Factory
  _____ ___ ___ ___ ___
 |     | . |   | . | . |
 |_|_|_|___|_|_|_  |___|
              |___|2.1.0

Connected to MongoDB v5.2.0 (at mongodb://localhost:27017)

Emma:
{
  "_id"   : ObjectId("57fe499fa81320d933000001"),
  "name"  : "Emma"
  "score" : 9,
}

Usage

MongoClient is the main entry point into Mongo, with MongoConnMgr underpinning everything. From there you can access all other components of MongoDB, namely MongoDb, MongoColl, MongoIdx.

MongoConnMgr                  : Connection Pool
  +-- MongoClient             : Client
       +-- MongoDb            : Database
            +-- MongoColl     : Collection
                 +-- MongoIdx : Index

Connecting

MongoClient may be created with a MongoConnMgr, which manages your connections to MongoDB for application wide, multi-threaded use:

connMgr := MongoConnMgr(`mongodb://localhost:27017`)
client  := MongoClient(connMgr)

When you create a MongoClient it immediately connects to MongoDB and verifies the version:

.    Fantom-Factory
 _____ ___ ___ ___ ___
|     | . |   | . | . |
|_|_|_|___|_|_|_  |___|
              |___|2.1.0

Connected to MongoDB v5.2.0

Note that MongoConnMgrPool will always query the supplied MongoDB host(s) to find the primary node, on which all read and write operations are performed.

Queries

Mongo works with BSON documents.

BSON documents are represented in Fantom as a Map of type [Str:Obj?]. All document keys must be strings. Document values can be any valid BSON type.

MongoDB stores BSON documents in Collections. Use the find() methods to query a collection. Using the friends database in the QuickStart Example we could do:

collection.findOne( ["name":"Emma"] )       // --> return the doc where 'name == Emma'
                                            //     ('Emma' must be unique)

collection.find                             // --> return ALL docs in the collection
collection.find( ["name":"Emma"] )          // --> return all docs where 'name == Emma'
collection.find( ["score": ["\$gt":7]] )    // --> return all docs with 'score > 7'

The $gt expression is an example of a Query operator.

This driver also supplies a little DSL for creating queries, in the form of MongoQ, which can make Mongo queries much easier to write.

// query using standard BSON
query := [
    "\$and" : [
        ["\$or": [["price": 0.99f], ["price": 1.99f]]],
        ["\$or": [["sale" : true ], ["qty"  : ["\$lt": 20]]]]
    ]
]

// same query using afMongo's DSL
query := MongoQ {
    and(
        or( eq("price", 0.99f), eq("price", 1.99f)  ),
        or( eq("sale", true),   lessThan("qty", 29) )
    )
}.query

collection.find(query)

To iterate over large result sets without loading it all into memory, use a Cursor. Cursors download documents in batches, behind the scenes, as and when required. Create and use a Cursor by using the find() method:

cursor := coll.find( ["score": ["\$gt":2]] )
cursor.batchSize = 10

while (cursor.isAlive)
    doc := cursor.next
    ...
}

cursor.kill

Write Commands

The insert() command is simple enough and is demonstrated in the QuickStart example.

update(), replace(), and delete() are similar in that they both take a query that describes which document(s) are to be updated / deleted. For most usages this will a simply be the id of the document in question:

collection.replace( ["_id": objId], [ ...new doc...] )

Mongo will throw a MongoErr should a write error occur.

ObjectId

All documents held in a collection need a unique id, held in a field named _id. If the _id field does not exist, MongoDB will create one for you of type ObjectId.

Note that _id does not need to an ObjectId, it can be any BSON type. It just needs to be unique in the collection.

Like marmite, people tend to have a love / hate relationship with the ObjectId. The good comments revolve around it having a natural sort that (roughly) corresponds to creation time. The bad is that it's a large human-unfriendly 24 char identifier, and in humongous collections it eats up precious bytes which means the index may not fit into RAM.

If sequential integers are more your thing when it comes to IDs, then MongoSeqs contains helper methods that use atomic updates on a named Collection to keep track of generated IDs.

Connection Fail Over

The MongoConnMgrPool accepts a replica set URL with multiple hosts (with optional ports):

mongodb://db1.example.net,db2.example.net:2500/?connectTimeoutMS=30000

When startup() is called, the hosts are queried to find the primary / master node. All read and write operations are then performed on this primary node.

When a connection to the master node is lost, all hosts are automatically re-queried to find a new master.

Remarks

The Fantom-Factory MongoDB driver was inspired by fantomongo by Liam Staskawicz.

If you're looking for cross-platform MongoDB GUI client then look no further than Robomongo / Robo 3T / Studio 3T Free!