Joining Multiple Indices’ Correlated Data in One ReactiveBase Component

[samstalner]

Hi there,

I was reading through the documentation, and was curious whether joining multiple indices’ data within a single ReactiveBase component is possible. I am aware that one can add multiple indices to the app prop (e.g. <ReactiveBase url='my_cluster_url' app='index1,index2'>), but I am interested in augmenting/merging search results such that the correlated information is available for each search result. For example, let’s say I have two indices Orders and Customers:

Orders index:

{
        …,
	customer:  {
		id: 123
	},
	orderDate: '12-15-2020',
	item: 'Yellow Shoe'
	…
}

Customers index:

{
	…,
	_id: 123,
	name: 'John Smith',
	dob: '01-02-1990',
	…
}

I now want to use ReactiveList to create a custom component that renders out the item and orderDate from Orders, as well as the corresponding customer’s name and dob from Customers (correlated on the id field). This would mean having a result object that contains a join of the fields from both the Orders and Customers index for that id.

Is there a way to do this natively using a ReactiveBase component? Are there examples of this somewhere?

Thank you so much!

[siddharthlatest]

There isn't a native way to do joins within ReactiveSearch. There are some workarounds though:

1. You can use transformResponse prop in ReactiveBase to query the other index and then stitch the data together. There's an example here that's showing that: https://docs.appbase.io/docs/reactivesearch/v3/overview/reactivebase/#props.
2. Another possible option is for you to change the data model within Elasticsearch: if you use the nested type to capture the customers within the order (or vice-versa), now you can query this more easily within ReactiveSearch.

[samstalner]

Thank you for your response @siddharthlatest! I was experimenting with the first option (using transformResponse) and just wanted to verify whether there was a cleaner way about it within ReactiveSearch. We will look into using the second option (creating a new index with all relevant information nested, and using the nestedField prop for queries). Thanks again!