Jun 28, 2022Backend Project Tutorial 2 - MongoDBThe second part to the PPC backend series, talking about using MongoDB.
Jasper Di Francesco6 min read

In this exercise, we will be building upon the backend server we created in the previous exercise - Creating a simple Express.js backend - but we will switch to using MongoDB to store data.

MongoDB is a noSQL database with a generious free tier. It is often used for small hobby projects, but is more than capable of managing data for huge applications.

Setting up

To start, follow the first 4 steps in this guide. Also on there you will find all the documentation for using MongoDB, we will be referring back to it a lot during this exercise.

Once you have created an account and launched an Atlas cluster, navigate to the database deployment dashboard. You should see your cluster there. It might still be launching but once it's done, click the Connect button and select Connect your application.

You should see a connection string there, mine looks like this: mongodb+srv://jasperdif:<password>@cluster0.3jv0w.mongodb.net/?retryWrites=true&w=majority. If you have something like that, congrats! We will come back to it later.

We will also be requiring the mongodb package for this, so navigate to the directory with your package.json and run the command npm install mongodb.

Connecting to MongoDB

Firstly import the MongoClient using the line.

venuesWatcher/index.js
const { MongoClient } = require('mongodb')

The reason we are wrapping this import in curly braces is because MongoClient is not the default export for mongodb. Read more about exports and modules here.

Next, we must delete our old database from our code, as in remove the database variable. We will replace it with the following few lines.

venuesWatcher/index.js
const url = 'YOUR_CONNECTION_STRING'
let client = new MongoClient(url)

Let's walk through this, we are getting our connection string from earlier, replacing the <password> section in it with our password and saving it as a constant value. Next we are instantiating a new MongoDB client with our connection string. This MongoClient is what will manage our connections to the database.

Now we want to change our app.listen functionality a little bit. We will be introducing an await statement so we firstly need to convert the function we pass to app.listen to be asynchronous. Then we want to call the connect method on our client. The listen function should look a little bit like this

venuesWatcher/index.js
app.listen(5000, async () => {
  await client.connect()
  console.log('Server started')
  console.log('Listening on port 5000')
})

Looks good! We have changed this function to be asynchronous, added in the await statement which just means we will stop execution until we have connected and we have actually connected to the database.

Now try running npm start. If there's a little bit of a delay and then you get the success messages, good job! You have successfully connected to your Atlas cluster.

Creating a database

The heirarchy of data in MongoDB is as such, clusters store databases which store collections which store documents. We have so far connected to the cluster, now we want to make a database inside that cluster.

Make two variables near the top of your index.js called venuesdb and venuescollection and set them to undefined. Surprise, they will hold our venues database and collection. After we have connected to our client in the app.listen function, we can create a new database and collection inside that database then save them to our variables with the following two lines:

venuesWatcher/index.js
venuesdb = client.db('venues')
venuescollection = db.collection('venues')

It is important to keep in mind that if you already had a database named 'venues' this will give you a connection to that, if not then it will create one.

Using MongoDB

Inserting

Inserting documents (some JSON data) into a MongoDB collection is incredibly easy. To do this, let's change our POST route to be the following:

venuesWatcher/index.js
app.post('/venues', async (req, res) => {
  await col.insertOne(req.body)
  res.status(200).send('Success')
})

Notice we have made 3 changes here, the function has become asynchronous so we put async before the parameters. Next we have swapped out Object.assign for the MongoDB insertOne function. The last important change is adding in .status(200) to our send line. This is us sending back a status code to the client, where the status code 200 indicates that the request was successful. You can read about all of the status codes here. It is very important you use status codes with all of your routes as it helps the client determine what type of response the server is giving.

Now that we have a more sophisticated database, we are going to change our data structure. Now, in our venues collection we will want to store venues in the following format:

{
  "name": "crumbs",
  "tags": ["cafe", "quick-eats"],
  "address": "405 High Street, Randwick NSW 2031",
  "notes": "Nice little cafe, affordable and excellent service.",
  "visits": 12,
  "rating": 4.5
}

So now, when we send a POST request to our server to add in a new cafe, it should have all of these fields in it.

Excellent, run your server and send a post request similar to the first exercise but with our new JSON format to add in another venue. You should get a success message on your console.

Finding

Finding documents can get a little bit trickier. First we establish a query object describing what we want to find in the database. Then we pass that query object to the findOne or find functions, depending on if we want all results that match returned or just one.

venuesWatcher/index.js
app.get('/venues', async (req, res) => {
  let query = { name: req.query.name }
  let item = await col.findOne(query)
  if (item) {
    res.status(200).send(item)
  } else {
    res.status(404).send({ error: 'Not found!' })
  }
})

Firstly make sure the GET request is asynchronous, else we can't use await. Then we want to build our query object by telling MongoDB "find things that have a name field the same as req.params.name". Then we collect the item from our database. findOne will return null if no documents matched the query, so we check to see if it found a matching document or not.

Try running your server again and searching for a venue with the name "crumbs". You should get back the example venue we added earlier. You might also notice an "id" field, MongoDB is very kind and automatically adds a unique ID to each document.

Good job! To sum up, now we have connected to our cluster, created a database, created a collection within that database and added some venues to it. We can also find a venue which matches a name parameter!

How might we write a route which would get all venues with a rating of 3? With a rating greater than 3? I'll leave these as an exercise to you, the MongoDB documentation might help.

In our next article we will explore CRUD operations for our database.