Converting Json String with Processor — Elasticsearch

Today we are going to use Json Processor to convert json string into json objects.

I will use two structures for the tests, the first will be a json string with an array.

json = [{\”color\”:\”red\”,\”value\”:\”#f00\”},{\”color\”:\”green\”,\”value\”:\ “#0f0\”}]

POST /_ingest/pipeline/_simulate?verbose=true
{
  "pipeline": {
    "description": "convert json string to document",
    "processors": [
      {
        "json": {
          "field": "json_string",
          "target_field": "colors"
        }
      }
    ]
  },
  "docs": [
    {
      "_index": "index",
      "_id": "id",
      "_source": {
        "json_string": "[{\"color\":\"red\",\"value\":\"#f00\"},{\"color\":\"green\",\"value\":\"#0f0\"}]"
      }
    }
  ]
}

I want the colors field to become the array with each json item.

Output:

"doc": {
            "_index": "index",
            "_id": "id",
            "_version": "-3",
            "_source": {
              "json_string": """[{"color":"red","value":"#f00"},{"color":"green","value":"#0f0"}]""",
              "colors": [
                {
                  "color": "red",
                  "value": "#f00"
                },
                {
                  "color": "green",
                  "value": "#0f0"
                }
              ]
            }

Now we will have another json: {\”employee\”:{\”name\”:\”sonoo\”,\”salary\”:56000,\”married\”:true}}.

We are going to make a change which is to add the “add_to_root” parameter, we want this “employee” object to be at the top level of the document. However, when using this parameter, we cannot define the “target_field”.

POST /_ingest/pipeline/_simulate?verbose=true
{
  "pipeline": {
    "description": "convert json string to document",
    "processors": [
      {
        "json": {
          "add_to_root": true, 
          "field": "json_string"
        }
      }
    ]
  },
  "docs": [
    {
      "_index": "index",
      "_id": "id",
      "_source": {
        "json_string": "{\"employee\":{\"name\":\"john doo\",\"salary\":1500,\"married\":false}}"
      }
    }
  ]
}

Output

          "doc": {
            "_index": "index",
            "_id": "id",
            "_version": "-3",
            "_source": {
              "employee": {
                "name": "john doo",
                "salary": 1500,
                "married": false
              },
              "json_string": """{"employee":{"name":"john doo","salary":1500,"married":false}}"""
            },
            "_ingest": {
              "pipeline": "_simulate_pipeline",
              "timestamp": "2023-01-27T04:12:05.8328892Z"
            }
          }

In both cases we were able to convert the json string. For more details on Json Processor parameters visit the documentation.

References

JSON processor | Elasticsearch Guide [8.6] | Elastic