Skip to main content

Create Fulltext Search Views with Nested Fields

This page explains how you can create fulltext search views for lower-level fields, also called nested fields. You can achieve this in two steps:

  1. Construct a JSON object, as outlined below.
  2. Use this JSON object to create a fulltext search view using the Create Search View API endpoint.

Step 1: Construct the JSON Object​

Your JSON object should have the following structure:

  • links (object):

    • [Collection Name] (string): Replace [Collection Name] with your collection's name.
      • analyzers (array of strings): Indexing analyzers for string values (default: ["identity"]).
      • fields (object): Defines attributes and nested structures for indexing.
      • includeAllFields (boolean): Whether to index all fields at a specific depth level (default: false).
      • trackListPositions (boolean): Whether list values should be indexed individually (default: false).
      • storeValues: Method to track attribute values. Options are id (to store values and enable EXISTS()) or none (default).

  • name (string): Name of the search view.

  • primarySort (array of strings): Default sorting order for the view.

  • type: Always set this to search.

JSON Example​

{
 "links": { 
   "testCollection": {            
     "analyzers": [ "text_en" ],
     "fields": {                
       "parties": { 
         "fields": {            
           "name": {} 
         } 
       } 
     },
     "includeAllFields": false,
     "storeValues": "none",
     "trackListPositions": false
    } 
  },
 "name": "testView",
 "primarySort": [],
 "type": "search" 
}

Accessing Nested Data​

You have two primary methods for accessing nested data.

Include All Fields​

Index all sub-fields of the primary field.

"includeAllFields": true

Specify Nested Field Paths​

This approach is more efficient, as it only indexes the specified fields. It's recommended if you know exactly which nested fields you want to index.

"fields": { 
  "topLevelField": {
    "fields": {
      "subLevelField": {}
    }
  }
}

If you anticipate adding more sub-fields later, then the first option provides more flexibility.

Step 2: Create the Fulltext Search View​

Access the endpoint using your preferred method.

Here's a cURL request example to create a search view using the previously constructed JSON:

curl -X 'POST' \
  'https://api-play.paas.macrometa.io/_fabric/<your-fabric-name>/_api/search/view' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -H 'Authorization: bearer <JWT_TOKEN>' \
  -d '<YOUR_JSON_OBJECT_HERE>'

Replace <YOUR_JSON_OBJECT_HERE> with the JSON you constructed in Step 1 and <JWT_TOKEN> with your JWT token.