How to split a json string value by character into some substrings in Apache Nifi

Question

i have a json file like the schema below and want to split the value of the key 'adress' by a character (for example a comma):

{
  "id": 123,
  "name": "James",
  "adress": "Oxford Street,21,London,England"
}

and convert it to the following (-> split the adress field by comma into these fields):

{
  "id": 123,
  "name": "James",
  "street": "Oxford Street",
  "house number": "21",
  "city": "London",
  "country": "England"
}

I found a solution in internet (http://ostack.cn/?qa=94733/), where someone splits their attibute into 2 key/values (with the jolt-processor) and it worked for me as well, but not for more than this split.

The processor could be 'Jolt Transform' or any with which i can edit json like in the schema above.

Thanks for the help, Lukas

Answer 1

You can use split function along with modify-overwrite-beta spec within JoltTransformJSON processor such as

[
  {
    "operation": "shift",
    "spec": {
      "@(0,adress)": "adr",
      "*": "&"
    }
  },
  {
    "operation": "modify-overwrite-beta",
    "spec": {
      "adr": "=split(',', @(1,&))",
      "street": "@(1,adr[0])",
      "house number": "@(1,adr[1])",
      "city": "@(1,adr[2])",
      "country": "@(1,adr[3])"
    }
  },
  {
    "operation": "remove",
    "spec": {
      "adr": "",
      "adress": ""
    }
  }
]

where the shift spec stands for generating the array(adr) to be processed through use of split, and remove is for deleting the unwanted key-value pairs.