Advanced Guides


Most filter components in ReactiveSearch provides a beforeValueChange prop. It is a callback function which accepts component’s future value as a parameter and returns a promise. It is called everytime before a component’s value changes. The promise, if and when resolved, triggers the execution of the component’s query and if rejected, kills the query execution. This method can act as a gatekeeper for query execution, since it only executes the query after the provided promise has been resolved.


Most of the time your needs can be solved using onValueChange. If you absolutely need beforeValueChange you should ensure that you don’t pass a function which takes a very long time to resolve the promise. The component goes in a locked state when using beforeValueChange and before the promise is resolved. This means all the state updates are suspended for the component.

Handling stream updates

The result components also allow streaming updates if you’re using to host your Elasticsearch cluster. You can enable this with the stream prop.

If you’re using streaming you can use render which receives an object with these values render({ data, streamData, .... }). The initial results from the query are received in the data key. The streamData parameter receives an array of objects when they’re created, deleted, or updated. If an object is updated, it contains a _updated key set to true. Similarly, if an object is deleted, it contains a _deleted key set to true. If an object is created, it contains neither of the two. This provides you with all the necessary information to handle streaming in your app suited to your needs. For example, we can utilize this to continuously handle streaming updates and merge new data with the existing:

render({ data, streamData }) {
    // generate an array of ids of streamResults
    const streamResultsIds = => item._id);

    return (
        // consider streamResults as the source of truth
        // first take existing data which is not present in stream data
        .filter(({ _id }) => !streamResultsIds.includes(_id))
        // then add data from stream data
        // remove data which is deleted in stream data
        .filter(data => !data._deleted)

Minimizing bundle size

ReactiveSearch from v2.3.1 also provides ES modules which can help you in reducing your app’s final bundle size. You can achieve this with tree shaking. If you’re unable to setup tree shaking in your project we recommend trying out the babel-plugin-direct-import.

yarn add -D babel-plugin-direct-import

After adding this you can update your .babelrc accordingly:

    "presets": [
    "plugins": [
              "name": "@appbaseio/reactivesearch",
              "indexFile": "@appbaseio/reactivesearch/lib/"

Now your import statements will only include the necessary modules. So,

import { ReactiveBase } from '@appbaseio/reactivesearch';

will include only the ReactiveBase module. Alternatively, you may avoid this step altogether and import using the full path, however the above method looks cleaner and you don’t have to worry about the component’s path in the library. Check out the example repo for the above setup.

Following also works with no extra setup, albeit a bit more explicit path:

import { ReactiveBase } from '@appbaseio/reactivesearch/lib/components/basic/ReactiveBase';


If you’re using create-react-app you might need to update your configurations if tree shaking is not working out of the box. You may try react-app-rewired instead of ejecting the app.