How to iterate over nested fields in painless script

Question

My document mapping looks like this:

                "_source": {
                    "id": 60,
                    "price": 60,
                    "locations": [
                      {
                         "id": 1
                         "price": 70
                      },
                      {
                         "id": 5,
                         "price": 80
                      }
                    ]
                }

So, price field can be both in root and in nested locations array. I want to be able to filter them by both values - if locationId is set in query parameters, then filter by price in locations, else - filter by root price. I wrote painless script:

            if(params['locationId'] != null){
                for (def location : doc['locations']) {
                    if (location.id == params['locationId']) {
                        if ((params['lte'] != null || location.price <= params['lte']) && (params['gte'] != null || location.price >= params['gte'])) {
                            return true;
                        } else {
                            return false;
                        }
                    }
                }
            }
            return (params['lte'] != null || params._source.price <= params['lte']) && (params['gte'] != null || params._source.price >= params['gte']);

And now I'm trying to use it:

            "query": {
                "bool": {
                    "filter": [
                        {
                            "script": {
                                "script": {
                                    "source": "my_script",
                                    "params": {
                                        "locationId": "2"
                                    }
                                }
                            }
                        }
                    ]
                }
            }

But I'm getting an error on second line of my script. No field found for [locations] in mapping with types [] How can I solve this? I also tried to access locations by params._source.locations, as I do similar in sort scripts, but params are not accessible from filters query

Answer 1

Answer #2. The bool Filter

Your query could be transformed into unscripted

Your filter condition is

[(locations.id == Id) AND (locations.price in range)] OR (price in range)

Unscripted bool query for the same mapping and document

GET /outer_inner_prices/_search?filter_path=hits.hits
{
    "query": {
        "bool": {
            "filter": [
                {
                    "bool": {
                        "should": [
                            {
                                "nested": {
                                    "path": "locations",
                                    "query": {
                                        "bool": {
                                            "must": [
                                                {
                                                    "term": {
                                                        "locations.id": {
                                                            "value": 5
                                                        }
                                                    }
                                                },
                                                {
                                                    "range": {
                                                        "locations.price": {
                                                            "gte": 1,
                                                            "lte": 80
                                                        }
                                                    }
                                                }
                                            ]
                                        }
                                    }
                                }
                            },
                            {
                                "range": {
                                    "price": {
                                        "gte": 1,
                                        "lte": 50
                                    }
                                }
                            }
                        ]
                    }
                }
            ]
        }
    }
}

Response

{
    "hits" : {
        "hits" : [
            {
                "_index" : "outer_inner_prices",
                "_type" : "_doc",
                "_id" : "1",
                "_score" : 2.0,
                "_source" : {
                    "id" : 60,
                    "price" : 60,
                    "locations" : [
                        {
                            "id" : 1,
                            "price" : 70
                        },
                        {
                            "id" : 5,
                            "price" : 80
                        }
                    ]
                }
            }
        ]
    }
}

My recommendation is to try to re-organize your data