How to update and strip the first character of an array field in elasticsearch

Question

I have the following within an index in elastic search listed as a json doc

Current

{
"bar": {
    "bar": [{
            "bar": [{
                    "bar": [{
                            "foo": "Y111111111111"
                        }
                    ]
        }
    ]
}

} }

Update needed

{
"bar": {
    "bar": [{
            "bar": [{
                    "bar": [{
                            "foo": "111111111111"
                        }
                    ]
        }
    ]
}

} }

How would I go about updating the index to strip the first character of the string where equal to bar?

I tried the following syntax which works for a single field but I receive an exception when running on an array field

{
  "error": {
  "root_cause": [
  {
    "type": "script_exception",
    "reason": "runtime error",
    "script_stack": [
      "ctx._source.foo = ctx._source.foo.substring(1);",
      "           ^---- HERE"
    ],
    "script": "ctx._source.foo = ctx._source.foo.substring(1);",
    "lang": "painless"
  }
 ],
 "type": "script_exception",
 "reason": "runtime error",
 "script_stack": [
  "ctx._source.foo = ctx._source.foo.substring(1);",
  "           ^---- HERE"
],
"script": "ctx._source.foo = ctx._source.foo.substring(1);",
"lang": "painless",
"caused_by": {
  "type": "illegal_argument_exception",
  "reason": "Illegal list shortcut value [foo]."
  }
 },
 "status": 400
}




POST test/_update_by_query 
{
"query": {
"prefix": {
  "foo": "Y"
}
},
"script": {
"source": "ctx._source.foo = ctx._source.foo.substring(1);"
}

Mapping

{
"TEST": {
    "mappings": {
        "properties": {
            "bar": {
                "properties": {
                "bar": {
                        "properties": {
                            "bar: {
                                "properties": {
                                "bar": {
                                        "properties": {
                                        "foo": {
                                                "type": "keyword"
                                            }
                                            }
                                    }
                                }
                            }
                        }
                    }
                   }
            }
        }
    }
    }

}

Answer 1

I would do it as follows. Find all documents whose foo field starts with Y and then update all the foo fields by stripping the first character:

POST test/_update_by_query
{
  "query": {
    "prefix": {
      "bar.bar.bar.bar.foo": "Y"
    }
  },
  "script": {
    "source": "ctx._source.bar.bar[0].bar[0].bar[0].foo = ctx._source.bar.bar[0].bar[0].bar[0].foo.substring(1);"
  }
}

PS: the query will depend on whether your bar fields are nested or not, but in case they are not, the above query should work.