# Build a Banking App Part 3: Methods of Fetching and Using Data ## Pre-Lecture Quiz [Pre-lecture quiz](.github/pre-lecture-quiz.md) ### Introduction At the core of every web application there's *data*. Data can take many forms, but its main purpose is always to display information to the user. With web apps becoming increasingly interactive and complex, how the user accesses and interacts with information is now a key part of web development. In this lesson, we'll see how to fetch data from a server asynchronously, and use this data to display information on a web page without reloading the HTML. ### Prerequisite You need to have built the [Login and Registration Form](../2-forms/README.md) part of the web app for this lesson. You also need to install [Node.js](https://nodejs.org) and [run the server API](../api/README.md) locally so you get account data. You can test that the server is running properly by executing this command in a terminal: ```sh curl http://localhost:5000/api # -> should return "Bank API v1.0.0" as a result ``` --- ## AJAX and data fetching Traditional web sites update the content displayed when the user click on a link or submit data using a form, by reloading the full HTML page. Every time new data needs to be loaded, the web server returns a brand new HTML page that needs to be processed by the browser, interrupting the current user action and limiting interactions during the reload. This workflow is also called a *Multi-Page Application* or *MPA*. ![Update workflow in a multi-page application](./images/mpa.png) When web applications started to become more complex and interactive, a new technique called [AJAX (Asynchronous JavaScript and XML)](https://en.wikipedia.org/wiki/Ajax_(programming)) emerged. This technique allows web apps to send and retrieve data from a server asynchronously using JavaScript, without having to reload the HTML page, resulting in faster updates and smoother user interactions. When new data is received from the server, the current HTML page can also be updated with JavaScript using the [DOM](https://developer.mozilla.org/en-US/docs/Web/API/Document_Object_Model) API. Over time, this approach has evolved into what we is now called a [*Single-Page Application* or *SPA*](https://en.wikipedia.org/wiki/Single-page_application). ![Update workflow in a single-page application](./images/spa.png) When AJAX was first introduced, the only API available to fetch data asynchronously was [`XMLHttpRequest`](https://developer.mozilla.org/en-US/docs/Web/API/XMLHttpRequest/Using_XMLHttpRequest). But modern browsers now also implement the more convenient and powerful [`Fetch` API](https://developer.mozilla.org/en-US/docs/Web/API/Fetch_API), which use promises and is better suited to manipulate JSON data. > While all modern browsers supports the `Fetch API`, if you want your web application to work on legacy or old browsers it's always a good idea to check the [compatibility table on caniuse.com](https://caniuse.com/fetch) first. ### Task In [the previous lesson](../2-forms/README.md) we implented the registration form to create an account. We'll now add code to login using an existing account, and fetch its data. Open the `app.js` file and add a new `login` function: ```js async function login() { const loginForm = document.getElementById('loginForm') const user = loginForm.user.value; } ``` Here we start by retrieving the form element with `getElementById()`, and then we get the username from the input with `loginForm.user.value`. Every form control can be accessed by its name (set in the HTML using the `name` attribute) as a property of the form. In a similar fashion to what we did for the registration, we'll create another function to perform a server request, but this time for retrieving the account data: ```js async function getAccount(user) { try { const response = await fetch('//localhost:5000/api/accounts/' + encodeURIComponent(user)); return await response.json(); } catch (error) { return { error: error.message || 'Unknown error' }; } } ``` We use the `fetch` API to request the data asynchronously from the server, but this time we don't need any extra parameters other than the URL to call, as we're only querying data. By default, `fetch` create a [`GET`](https://developer.mozilla.org/en-US/docs/Web/HTTP/Methods/GET) HTTP request, which is what we are seeking here. ✅ `encodeURIComponent()` is a function that escapes special characters for URL. What possible issues could we possibly have if we do not call this function and use directly the `user` value in the URL? Let's now update our `login` function to use `getAccount`: ```js async function login() { const loginForm = document.getElementById('loginForm') const user = loginForm.user.value; const data = await getAccount(user); if (data.error) { return console.log('loginError', data.error); } account = data; navigate('/dashboard'); } ``` First, as `getAccount` is an asynchronous function we need to match it with the `await` keyword to wait for the server result. As with any server request, we also have to deal with errors cases. For now we'll only add a log message to display the error, and come back to it layer. Then we have to store the data somewhere so we can later use it to display the dashboard informations. Since we the `account` variable does not exist yet, we'll create a global variable for it at the top of our file: ```js let account = null; ``` After the user data is saved into a variable we can navigate from the *login* page to the *dashboard* using the `navigate()` function we already have. Finally, we need to call our `login` function when the login form is submitted by modifying the HTML: ```html