Elasticsearch Java api split a field

Elastic Stack Elasticsearch
1

Hi, I am trying to execute an aggregation to get the distinct"roles"
the field "roles" wich contain different roles separated with comma here an example:

          "roles": ["\"ROLE_CUSTOMERGROUP,ROLE_PRONTO,western service\""]

the problem if i use a standard aggregation on the field roles, elasticsearch will split " western service " :

"aggregations": {
"roles": {
  "doc_count_error_upper_bound": 0,
  "sum_other_doc_count": 0,
  "buckets": [
    {
      "key": "role_customergroup",
      "doc_count": 3124
    },
               {
      "key": "role_pronto",
      "doc_count": 1945
    },
    {
      "key": "role_rustic",
      "doc_count": 1179
    },
    {
      "key": "role_services",
      "doc_count": 1179
    },
    {
      "key": "services",
      "doc_count": 1179
    },
    {
      "key": "west",
      "doc_count": 1179
    }
  ]
}

is there any solotion to this problem? should i use a script?

Thank you !

2

This is because of the analyzer, if you don't want elastic search to "break" the words on that field you should set "roles" field to be not_analyzed.