Skip to main content
Version: 2021fa

Lecture 3

Lecture Slides

Assignment 2 (due 10/21 6:29 PM on CMS)

Install Postman

Before the lecture

Create tsconfig.json

From now on, we will be using a tsconfig.json file within every Node project we create (a recap on how to do that is below this section). Essentially, the tsconfig.json is a file at the root of a Node project which indicates it is using TypeScript, and allows us to configure the TypeScript compiler. If you'd like to follow the lecture synchronously, you can put the following chunk of code into the root directory of your lecture 3 project. If you're more curious about how the file works, you can refer to this link.

"compilerOptions": {
"target": "ES2021",
"outDir": "dist",
"lib": ["dom", "dom.iterable", "esnext"],
"baseUrl": "./src",
"skipLibCheck": true,
"esModuleInterop": true,
"allowSyntheticDefaultImports": true,
"strict": true,
"forceConsistentCasingInFileNames": true,
"isolatedModules": true,
"noUnusedLocals": true,
"importsNotUsedAsValues": "error",
"module": "commonjs",
"moduleResolution": "node",
"resolveJsonModule": true
"include": ["src"]

Install Postman

Use the link above to install Postman.

Postman and Request Bodies


Instead of always going to the endpoint in the browser, a robust way of testing our endpoints is to use Postman.

Postman is a software that allows you to simulate requests that could be sent by a user to your backend. It is useful for testing and ensuring that the behavior of your requests (including necessary headers) is what you expect.

POST Request

Usually when you want to send a POST request you also want to send information with it. Situationally, you want to do this using request bodies rather than query parameters.

Let's say we have this addSong endpoint:

index.ts'/addSong', (req, res) => {
const song: Song = { name:, rating: req.body.rating };
res.send(`Song ${} added!`);

where Song is the type:

type Song = {
name: string;
rating: number;

Previously, we may have considered using query parameters for sending data for the backend. There's nothing wrong with that; for example, we could have used /addSong?name=Despacito&rating=5. However, this can lead to extremely long URLs, and limit us from sending more complicated data. That's where request bodies come in handy. We can instead send request data in JSON format to the backend, allowing us to use the data more easily and integrate it seamlessly with our backend (which happens to be in TypeScript, so we can easily deal with it).

The snippet above tells express to listen for POST requests at endpoint /addSong. req.body is a JavaScript object, and we access its properties and req.body.rating to add a new song to our array of songs.

Now, we should have a working POST endpoint that does something with the request body.

However, we can't test request bodies quite as easily through the browser; we can check that this endpoint is working using Postman. Set the request type to POST and URL as localhost:8080/addSong. To send a request body, first go to Headers and add a new key Content-Type with value application/json. This says we are sending JSON input (essentially, an object or dictionary) in our request body. JSON is generally used in POST requests to send a payload (and also for more nested structures). In the Body section, select the raw radio button and enter the following in the text field:

"name": "Despacito",
"rating": 5

We will be sending name with a value of "Despacito" and rating with a value of 5 in the request body.

Sending this request, you should see the corresponding song printed out to the console by the endpoint.

Now, let's create another POST endpoint to update a song's rating. This will also use a request body with just a name field, which should match the song we want to update.

index.ts'/updateRating', (req, res) => {
for (const song of songs) {
if ( === {
song.rating = req.body.rating;
res.send('Rating updated!');

DELETE Request

When creating APIs, we use the DELETE request method to quite simply delete a specific resource. This should be pretty straightforward: we simply take the name of the song to delete through the request body, and create a new version of the songs without the specified song. We then send text to the requester that it was deleted.

app.delete('/removeSong', (req, res) => {
const newSongs = [];
for (let song of songs) {
if ( !== {
songs = newSongs;
res.send(`Song ${} deleted!`);

And with that, we're done!

Intro to Databases and Firebase

The song API we just made "works": we can add songs and then get them while running the Express server. But it has one fatal flaw: try stopping the server and then running it again. You'll see that all the music is gone! We need some kind of persistent storage for this data: throughβ€”you guessed itβ€”a database.

Why do we need a database for our backend?

  • Data stored within Node.js is per instance
  • Most applications require persistence
  • Data analysis
  • Performant data location
  • Offloading unneeded data from memory

MySQL + Relational Databases

  • Stores data in tables, utilizing rows and tables.
  • Is relational (think a tuple)
  • Has a schema

NoSQL and Firestore

We will focus on NoSQL

  • Many NoSQL implementations are schema-less or have a partial schema
  • Firestore is a cloud-hosted NoSQL database
  • Very flexible and can be used with most popular languages
  • Uses documents to store data
  • Efficient querying for data


  • SQL databases have a predefined schema, whereas NoSQL databases can abide to any structure you want it to.
  • NoSQL databases are better suited for large sets of data, but not for complex queries.
  • SQL databases tend to be less expensive for smaller datasets, but also less flexible.
  • SQL leans towards strong consistency whereas NoSQL favors eventual consistency (i.e. there may be some delay in getting the response back)
  • SQL databases tend to be vertically scalable (need more computing power on one machine to store more data) while NoSQL databases tend to be horizontally scalable (can distribute storage and compute power on multiple machines)
  • Examples of SQL databases: MySQL, PostgreSQL
  • Examples of NoSQL databases: Firebase, MongoDB

What is Firebase?

  • Firebase is a Backend as a Service (BaaS) offered by Google
    • Allows you to store data
    • Host websites
    • Authentication
  • NoSQL DB
    • Not only SQL
    • Non relational

Why Firebase?

  • Real-time operations
  • Firebase Authentication
  • Built-in analytics
  • Also supports hosting/deployment
  • Integration with other Google services
  • Structure we’re familiar with!

Basic Database Manipulations

People usually call that CRUD, which stands for:

  • Create/Insert - Create a document (will fail if the document exists)
  • Read/Find/Query - To search for documents based on their properties
  • Update - Update an existing document (will fail otherwise)
  • Delete - Delete an existing document

For convenience, most NoSQL database also provides an upsert operation. It will create the document or update an existing document. You can think of that as an atomic operation that does:

if (document.exists()) {
} else {

In Firestore, you can either insert a new document with a specified ID, or allow Firestore to generate its own ID for you.

The update method in Firestore allows you to update certain fields of the document without overwriting the entire thing.

Sample code

The following code demonstrates how we can do basic CRUD with Firestore. Note that the code below does not care about what are the fields of a post, because Firestore doesn't require you to have a predefined set of fields. This gives you flexibility when writing your backend code.

import admin from 'firebase-admin';
import express from 'express';

// require the service account: note the file path
const serviceAccount = require('../service-account.json');
// initialize the firebase app
credential: admin.credential.cert(serviceAccount),

const db = admin.firestore();
const app = express();
const port = 8080;
// allow request body parsing

// check connections
app.get('/', (_, res) => {

// create a post type and post with id
type Post = {
content: string;
name: string;

type PostWithID = Post & {
id: string;

// CRUD with firestore
let posts1: Post[] = [{ content: 'I miss wellness days', name: 'Becky' }];

// posts collection from db
const postsCollection = db.collection('posts');

// GET requests: get the songs
app.get('/getPostsLocal', (_, res) => {

// use firebase instead
app.get('/getPostsFirebase', async (_, res) => {
const postsSnapshot = await postsCollection.get();
const allPostsDoc =;
const posts: PostWithID[] = [];
for (let doc of allPostsDoc) {
const post: PostWithID = as PostWithID; =;

// read posts by name
app.get('/getPosts/:name', async function (req, res) {
const name =;
const postsSnapshot = await postsCollection.where('name', '==', name).get();
const allPostsDoc =;
const posts: PostWithID[] = [];
for (let doc of allPostsDoc) {
const post: PostWithID = as PostWithID; =;

// read posts by id
app.get('/getPostById/:id', async function (req, res) {
const id =;
const postsSnapshot = await postsCollection.doc(id).get();
const post: PostWithID = as PostWithID;

// sort posts in descending order by name
app.get('/getPostsSorted', async function (req, res) {
const postsSnapshot = await postsCollection.orderBy('name', 'desc').get();
const allPostsDoc =;
const posts: PostWithID[] = [];
for (let doc of allPostsDoc) {
const post: PostWithID = as PostWithID; =;

// POST method: create a new post'/addPostLocal', (req, res) => {
const post: Post = req.body;
res.send(`Post created by ${}!`);

// generate a document with a random name to store the post's data'/addPostFirebase', async function (req, res) {
const post: Post = req.body;
const postDoc = postsCollection.doc();
await postDoc.set(post);

// POST method: update an existing post'/updatePostLocal', (req, res) => {
for (let post of posts1) {
if ( === {
post.content = req.body.content;
res.send('content updated!');

// update by id'/updatePostFirebase/:id', async function (req, res) {
const newPost: Post = req.body;
const id: string =;
await postsCollection.doc(id).update(newPost);

// DELETE methdod: delete a post
app.delete('/removePostLocal', (req, res) => {
const newPosts = [];
for (let post of posts1) {
if ( !== {
posts1 = newPosts;
res.send(`Post by ${} deleted!`);

// delete by id
app.delete('/removePostFirebase/:id', async function (req, res) {
const id =;
await postsCollection.doc(id).delete();

app.listen(port, () => console.log(`App started on port ${port}!`));