Skip to content

Running search queries via REST API

Search queries follow the capabilities that OpenSearch provides. However, the request schema is different from the one that OpenSearch uses.

Basic query structure

The search REST API provides an OpenSearch-compatible search syntax with some adjustments. This allows developers to use the capabilities of OpenSearch.

The following example shows the basic structure of a search query:

{
    "query": {
        "must": [],
        "should": [],
        "must_not": [],
        "filter": [],
    },
    "aggregations": {},
    "page": 0,
    "pageSize": 30,
    "scope": [],
    "sort": []
}

The following table provides more information about each property in the search query:

Property Description
query Same syntax as an OpenSearch Boolean Compound Query.
aggregations Same syntax as OpenSearch Aggregations.
page Pagination index. Default is 0.
pageSize Results per page. Default is 100.
scope A list of content source IDs, to which the search request is limited to. If omitted or empty, all content sources are searched.
sort Same syntax as OpenSearch Sort.

The filter section of the query is always enriched during processing with a filter for the user's ACLs. This ensures a search result only contains the allowed fields.

Note that a search query is currently limited to 10000 results.

Example queries

This section provides sample search queries for several scenarios.

Search for a title in WCM

{
  "query": {
    "must": [
      {
        "match": {
          "documentObject.title": "Product"
        }
      }
    ]
  }
}

This search query returns all documents containing the word Product in their title field.

Search for a collection in DAM

{
  "query": {
    "must": [
      {
        "wildcard": {
          "documentObject.name": {
            "value": "*Product*",
            "case_insensitive": true
          }
        }
      },
      {
        "match": {
          "documentObject.type": "collection"
        }
      }
    ]
  },
  "page": 0,
  "pageSize": 10
}

This search query returns all documents containing the word Product in their name field and the type is collection.

Search for an asset in DAM

{
  "query": {
    "must": [
      {
        "wildcard": {
          "documentObject.name": {
            "value": "*Product*",
            "case_insensitive": true
          }
        }
      },
      {
        "match": {
          "documentObject.type": "asset"
        }
      }
    ]
  },
  "page": 0,
  "pageSize": 10
}

This search query returns all documents containing the word Product in their name field and the type is asset.

Search using a query string

{
  "query": {
    "must": [
       {
         "query_string": {
           "query": "Products AND (01 OR 03)"
         }
       }
    ]
  }
}

You can use query strings to define a search query. This search query returns all documents containing the word Products and the number 01 or 03.

Limiting scope to one content source

{
  "query": {
    "must": [
      {
        "match": {
          "documentObject.title": "Product"
        }
      }
    ]
  },
  "scope": [
    "619b9540-b2aa-45e6-a69b-0a0ab6799a78"
  ]
}

This returns all documents containing the word Product in their title field but only for the content source with ID 619b9540-b2aa-45e6-a69b-0a0ab6799a78.

Sorting results by a field

{
  "query": {
    "must": [
      {
        "match": {
          "documentObject.title": "Product"
        }
      }
    ]
  },
  "sort": [
    {
      "documentObject.published": {
        "order": "desc"
      }
    }
  ]
}

This returns all documents containing the word Product in their title field but they are sorted by their published date in descending order.

Using filters

{
  "query": {
    "must": [
      {
        "match": {
          "documentObject.title": "Product"
        }
      }
    ],
    "filter": [
        {
            "term": {
                "documentObject.locale": "fr"
            }
        }
    ]
  }
}

This returns all documents containing the word Product in their title field but only the ones with the locale being fr. The filter has no influence on scoring and is purely binary.

Related information