MacLochlainns Weblog

Sometimes a bit of humor helps with a new topic. Creating a database in MongoDB is a two-step process, like the Texas Two-Step (a nick name for a country/western two-step danced in common time). A Texas Two Step is a one-two, one-two shuffle which is like the two-step process for how you create a MongoDB database.

While databases in MongoDB are a multiuser sandbox like a relational database, you can’t simply create them and grant them privileges. You must first USE the database and then put a collection (a.k.a., equivalent to a table in MySQL) in it. This blog post shows you how to create a MongoDB play database with an actor collection. It shows you the commands to create the a database with one collection in it and how to verify its existence.

You connect through the mongo shell, like:

mongo

The first-step requires you to connect to the play database even though it doesn’t exist. You do that by typing:

> use play

You can verify you’re in the play database by typing the db command, which the mongo shell treats as an expression. It returns play, as the name of the current database.

However, when you call the shell show dbs helper:

> show dbs

It displays:

admin   0.000GB
config  0.000GB
local   0.000GB

You also should note that the play database still doesn’t exist when you run the JavaScript equivalent to the shell show dbs helper:

> db.getMongo().getDBs()

It returns the following list of databases, which excludes the as yet not created play database:

{
  "databases" : [
    {
      "name" : "admin",
      "sizeOnDisk" : 32768,
      "empty" : false
    },
    {
      "name" : "config",
      "sizeOnDisk" : 61440,
      "empty" : false
    },
    {
      "name" : "local",
      "sizeOnDisk" : 81920,
      "empty" : false
    }
  ],
  "totalSize" : 176128,
  "ok" : 1
}

You create an actor collection (a.k.a. the equivalent of a relational table) with the following syntax, where db maps to the play database that you’re using. Note that you must create or insert one document to begin a document collection. While you can use the insert method, you should use either the newer insertOne() or insertMany() methods.

> db.actors.insertOne({"actor" : {"first_name" : "Chris", "last_name" : "Pine"}, "age" : 40 })

A new call to the shell show dbs helper:

> show dbs

displays your new play database:

admin   0.000GB
config  0.000GB
local   0.000GB
play    0.000GB

You can add two more documents (a.k.a. for rows in a relational database) with the insertMany() method:

> db.actors.insertMany([{"actor" : {"first_name" : "Chris", "last_name" : "Evans"}, "age" : 39 }
                       ,{"actor" : {"first_name" : "Chris", "last_name" : "Pratt"}, "age" : 41 }])

A quick word to the JavaScript novices out there. Don’t forget the square brackets ([{},{},...]) in the insertMany() method call or you’ll get an error like this:

2021-04-12T16:20:13.237-0600 E QUERY    [js] TypeError: documents.map is not a function :
DBCollection.prototype.insertMany@src/mongo/shell/crud_api.js:295:1
@(shell):1:1

You call the shell show collections helper or db.getMongo().getCollectionNames() JavaScript function to display the collections in the play database. The show collections displays a list of collections, and the db.getMongo().getCollectionNames() displays an JavaScript array of collections.

If you’re like me, Mongo’s db convention is a bit risky that I could do something in the wrong database. So, I put the following function into my .mongorc.js (a.k.a., MongoDB Resource file):

prompt = function() { var dbName = db; return dbName + "> " }

It ensures you will see the current database name to the left of the prompt (“>”), like:

play>

You can query the documents from the actors collection with the following:

play> db.actors.find().pretty()

It returns:

{
  "_id" : ObjectId("6074c692813c5a85db9cc9df"),
  "actor" : {
    "first_name" : "Chris",
    "last_name" : "Pine"
  },
  "age" : 40
}
{
  "_id" : ObjectId("6074c7ea813c5a85db9cc9e0"),
  "actor" : {
    "first_name" : "Chris",
    "last_name" : "Evans"
  },
  "age" : 39
}
{
  "_id" : ObjectId("6074c7ea813c5a85db9cc9e1"),
  "actor" : {
    "first_name" : "Chris",
    "last_name" : "Pratt"
  },
  "age" : 41
}

As always, I hope this helps those looking for a solution.