Second, a form of local storage must be used as the main data source of the application. This storage must be continuously synchronized with a remote data source. Luckily, there are databases like PouchDB, which store data locally and can be synced automatically with a CouchDB database in a server.
In this tutorial, we’re going to build an anonymous chat app with offline capabilities. It will look like this:
We’re going to use Create React App to automatically generate a progressive web app with React and RxDB, a reactive, client-side, and offline-first database that will sync with a server-side CouchDB database.
You will need:
- Node version 6 or higher. You can use nvm to switch between Node versions.
- A modern browser that supports IndexedDB.
For reference, you can find the entire source code of on this GitHub repository:
GitHub – eh3rrera/react-rxdb-example: Anonymous chat app built with React, RxDB and PouchDB that works offline.
Anonymous chat app built with React, RxDB and PouchDB that works offline. – GitHub – eh3rrera/react-rxdb-example: Anonymous chat app built with React, RxDB and PouchDB that works offline.
Setting up the React app
Install (or upgrade to the latest version)
npm install -g create-react-app
Then, create a new app and
cd into it:
create-react-app offline-anonymous-chat cd offline-anonymous-chat
Now let’s install all the dependencies for our project by executing:
npm install --save concurrently moment pouchdb-adapter-http pouchdb-adapter-idb pouchdb-server react-toastify rxdb rxjs serve
Here’s the description of each dependency:
- concurrently. We’ll need it to run two npm scripts at the same time.
- moment. We’ll need it to format the creation date of the message.
- pouchdb-adapter-http. PouchDB adapter for communicating with an external CouchDB (or CouchDB-like) database.
- pouchdb-adapter-idb. PouchDB adapter to use IndexedDB in a browser. You can use PouchDB (and RxDB) in different environments by just switching the adapter.
- pouchdb-server. This will work as our server-side database.
- react-toastify. We’ll need it to show notifications in the app for the database events.
- rxdb. This will work as our client-side database.
- rxjs. RxDB handles data in a reactive way, so it depends on rxjs.
- serve. When testing the offline functionality, we’ll need an HTTP server to serve our app (more on this later).
Next, edit the
start script in the
package.json file. Instead of:
"start": "react-scripts start",
It should be:
"start": "concurrently "npm run server" "react-scripts start"",
Also, we should add the
"server": "pouchdb-server -d ./db",
This will execute the PouchDB server at the same time the application is running. The database file will be stored in the
db directory. You can find all the configuration options for the database server here.
When you execute
npm start, you can go to http://localhost:3000 (if it’s not opened automatically) to view the app generated by
As you can see, there are just two databases created that are used internally by PouchDB.
In the next section, we’ll create the schema for our app database.
Creating the database schema
We can use schemas to define how the data will look and validate every inserted document. The data is organized into collections and every collection has its own schema.
For our app, we are going to use a collection named
messages which will store the date the message was created and the message itself.
In the above schema:
- The version number is zero. If the version is greater than zero, you have to provide a data migration strategy.
- It has two properties of type
message. The first one is the primary key, and it represents the date the message was inserted as the number of milliseconds since the Unix epoch.
- The property
You can learn more about schemas here.
Now let’s configure the local database.
Configuring the local database
Add the following imports to
In addition to importing the main RxDB module and the schema created in the previous section, we’re importing the module
As RxDB is a reactive database, you can subscribe to queries to receive new results in real-time, but executing a query every time this happens can impact performance, so the option
QueryChangeDetector optimizes observed queries by getting new results from database events instead.
This option is currently in beta and disabled by default, but we can enable it with the following code:
We can also enable debugging so every time an optimization is made, a message in the console is shown.
Next, configure the adapters to use IndexDB as the storage engine, and enable syncing to a remote database over HTTP:
And declare as constants the URL of the remote database and the name of the local database:
Inside the class, let’s declare a method to create the database:
The database is created in an asynchronous way, but instead of using callbacks or promises, we’re going to use async/await for more concise and clean code. Here’s a good article comparing the three approaches of writing asynchronous code.
This way, we create the database passing a name, and adapter, and a password (which must have at least 8 characters):
Next, we enable the leader-election algorithm which makes sure that always exactly one tab is managing the remote data access (in case there are multiple tabs of the application at the same time). When the leader is elected, a crown will be shown next to the page’s title:
Then, we create the
messages collection passing the schema:
More great articles from LogRocket:
- Don't miss a moment with The Replay, a curated newsletter from LogRocket
- Learn how LogRocket's Galileo cuts through the noise to proactively resolve issues in your app
- Use React's useEffect to optimize your application's performance
- Switch between multiple versions of Node
- Discover how to animate your React app with AnimXYZ
- Explore Tauri, a new framework for building binaries
- Compare NestJS vs. Express.js
Finally, let’s set up the replication feature and return the database object:
Now let’s integrate this method into our React app.
Building the app
Let’s start by adding the imports for the component to show notifications and the moment library to handle dates:
Next, add the following constructor to the class:
We have two properties as the state of our component, one for new messages, and an array to store all the messages.
We also defined an array to store the handlers of the subscriptions we’re going to use so we can unsubscribe when necessary, along with the binding of the functions to add a new message and handle the text box’s changes for new messages.
componentDidMount, let’s call the method
createDatabase, and then subscribe to a query that returns all the messages sorted by
$() will return a RxJS observable which streams every change to the data returned by the query. If any messages are returned, a notification is shown and the state is changed.
Notice that we also add a reference to the subscription to the
subs array. This way, when the component is about to be unmounted, we’ll use those references to unsubscribe from the observables:
Next, modify the
render method so it looks like this:
As you can see, the rendering logic is simple.
renderMessages converts each message object into a
div element showing the relative insertion date of the message with the help of the
moment library (remember that the
id represents the date in milliseconds) and the message:
This is the definition of the method
And to add a message, we create an object and insert it into the database in the following way, setting the state to an empty string for the next message:
Finally, probably is a good idea to show the replication events to understand what is going on. The method
RxCollection.sync() returns a RxReplicationState object that can be used to observe events and to cancel the replication.
Modify the last part of the method
createDatabase so it looks like this:
Some of these events emit an object, that’s why in some cases they are printed in the console in addition or instead of being shown as notifications. Here’s a description of each event:
change$. Triggered every time a document is replicated. It can emit a boolean or an object describing the change.
docs$. Emits each replicated document.
falsedepending on whether the replication is running or not.
falsedepending on whether the replication is completed or not (actually, only one-time-replications will complete).
error$. Triggered when an error occurs during the replication process.
Also, we added these subscriptions to the
subs array so we can unsubscribe when necessary.
And that’s it. Let’s test the app.
Testing the app
npm start and go to
http://localhost:3000. You should see something like this:
Start adding some messages. Notice the crown in the title, it indicates this tab is the leader. If you open another tab, and close the first one, now the other should have the crown:
Thanks to the reactive nature of RxDB (and the fact that all the tabs in a browser share the same local storage), the database state is broadcasted in real-time, which means that the changes in one tab will be reflected automatically in another tab:
Go to http://localhost:5984/_utils and now you should see the app’s database:
Click on the database name and then on the pencil icon of one of the documents to edit it:
The change made on the database web interface should be applied to the app almost instantly:
So it seems to be working fine, but does it work in offline mode?
The production build of the app generated by
create-react-app is a fully functional, offline-first progressive web app.
There are some considerations we must take into account, but basically, we just have to build the application and serve it. Execute:
npm run build
This command creates a
build directory with a production build of your app.
Now add the following scripts to the
"http": "serve -p 3000 -s build", "offline": "concurrently "npm run server" "npm run http""
npm run offline
Since the app now will install a service worker and a manifest, probably you may want to test it using your browser’s incognito-mode, to avoid conflicts if you want to go back to the development version later.
Once again, go to
http://localhost:3000, the application will install the service worker and the browser will cache the files. If you’re using Chrome, you can check this by going to Developer Tools — Application and choose the sections Manifest and/or Service Workers:
Now go offline by either choosing the Offline option from the Service Worker panel, the Offline option of Developer Tools’ Network tab, or just disconnecting from the network.
When you reload the app, you should still be able to insert messages:
After inserting some messages, go back online. After a while, you should see some notifications about replication:
If you go to the database web interface, the database should contain the new documents:
LogRocket combines session replay, product analytics, and error tracking – empowering software teams to create the ideal web and mobile product experience. What does that mean for you?
Instead of guessing why errors happen, or asking users for screenshots and log dumps, LogRocket lets you replay problems as if they happened in your own browser to quickly understand what went wrong.
No more noisy alerting. Smart error tracking lets you triage and categorize issues, then learns from this. Get notified of impactful user issues, not false positives. Less alerts, way more useful signal.
The LogRocket Redux middleware package adds an extra layer of visibility into your user sessions. LogRocket logs all actions and state from your Redux stores.
Modernize how you debug your React apps — start monitoring for free.