"fields" ignored on MLT query with artificial document

Elastic Stack Elasticsearch
1

Hi everyone!
When I execute an MLT query with an artificial document the "fields" field is ignored and all the fields from the artificial document are used for scoring.
For instance I built an index with these settings:

PUT /myindex
{
    "mappings": {
        "properties": {
            "body": {
                "type": "text",
                "analyzer": "rebuilt_italian",
                "term_vector": "yes"
            },
            "url": {
                "type": "keyword"
            },
            "title": {
                "type": "text",
                "analyzer": "rebuilt_italian",
                "term_vector": "yes"
            }
        }
    },
    "settings": {
        "index": {
            "analysis": {
                "filter": {
                    "italian_stemmer": {
                        "type": "stemmer",
                        "language": "italian"
                    },
                    "italian_elision": {
                        "type": "elision",
                        "articles": [
                            "c",
                            "l",
                            "all",
                            "dall",
                            "dell",
                            "nell",
                            "sull",
                            "coll",
                            "pell",
                            "gl",
                            "agl",
                            "dagl",
                            "degl",
                            "negl",
                            "sugl",
                            "un",
                            "m",
                            "t",
                            "s",
                            "v",
                            "d"
                        ],
                        "articles_case": "true"
                    },
                    "italian_stop": {
                        "type": "stop",
                        "stopwords": "_italian_"
                    }
                },
                "analyzer": {
                    "rebuilt_italian": {
                        "filter": [
                            "italian_elision",
                            "lowercase",
                            "italian_stop",
                            "italian_stemmer"
                        ],
                        "tokenizer": "standard"
                    }
                }
            }
        }
    }
}

And then I add this document:

PUT /myindex/_doc/1
{
    "title": "New document title",
    "body": "New document body",
    "url": "http://example.com"
}

When I explain the query with these parameters:

POST /myindex/_explain/1
{
    "query": {
        "more_like_this": {
            "fields": [
                "title",
                "url"
            ],
            "like": [
                {
                    "doc": {
                        "title": "Document title",
                        "body": "Document body",
                        "url": "https://example.com"
                    }
                }
            ],
            "min_term_freq": 1,
            "min_doc_freq": 1,
            "max_query_terms": 12
        }
    }
}

I can see that the "body" field is used for scoring even if I didn't set it in the "fields".
Is this the correct behavior?

Cheers,
Mirco

edit: I'm using the version 7.15.0 of Elasticsearch

2

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.