breaking field value on hypen

Question

I have following sample docs in Elasticsearch

[
  {
    "_index": "beatbox-2018.11.19",
    "_source": {
      "connection": "10",
      "user": "op-dashboard",
      "key": "monolith_connection_sniffer"
    }
  },
  {
    "_index": "beatbox-2018.11.19",
    "_source": {
      "connection": "10",
      "user": "op-dashboard",
      "key": "monolith_connection_sniffer"
    }
  }
]

When I query on user, I got expected result.

curl -X GET \
  'http://127.0.0.1:9200/beatbox-2018.11.19/_search?q=user:op-dashboard'

In Grafana:

I was trying to add some query with Variable for user field.

{   "find": "terms",   "field": "user" }

But I got tokenised values of user field.

`op`, `dashboard`

In background, following payload is sending for query

{
  "size": 0,
  "query": {
    "bool": {
      "filter": [
        {
          "query_string": {
            "analyze_wildcard": true,
            "query": "*"
          }
        }
      ]
    }
  },
  "aggs": {
    "1": {
      "terms": {
        "field": "user",
        "size": 500,
        "order": {
          "_term": "asc"
        }
      }
    }
  }
}

Query returns tokenised result. How can I stop it?

I have already tried with following template

{
  "index_patterns": [
    "beatbox*"
  ],
  "mappings": {
    "doc": 
      "properties": {
        "user": {
          "type": "text",
          "fielddata": true,
          "analyzer":"whitespace",
          "search_analyzer": "whitespace"
        }
      }
    }
  }
}

And also with analyzer

{
  "index": {
    "analysis": {
      "default": {
        "analyzer": {
          "analyzer_keyword": {
            "tokenizer": "whitespace"
          }
        }
      }
    }
  }
}

Mapping for index:

{
   "beatbox-2018.11.19":{
      "mappings":{
         "doc":{
            "_all":{
               "enabled":false
            },
            "numeric_detection":true,
            "properties":{
               "connection":{
                  "type":"long"
               },
               "key":{
                  "type":"text",
                  "norms":false,
                  "index_options":"freqs"
               },
               "user":{
                  "type":"text",
                  "fielddata":true
               }
            }
         }
      }
   }
}

Any help?

Answer 1

You should you Keyword Datatype of elasticsearch instead of text datatype in user fields as you are aggregating it.