CODEWITHSUNDEEP
elasticsearch
Vortex Commerce
Vortex Commerce

Reputation: 83

Elastic Search Document Boosting (like Solr elevateIds)

Looking to move from Solr to Elastic Search due to its amazon REST API and suitability for clustering. However I am struggling to figure out a way to boost at query time specific documents like in Solrs elevate method. Is there a way to do this and if so any examples would be really helpful :)

Upvotes: 2

Views: 1246

Answers (1)

John Petrone
John Petrone

Reputation: 27507

I assume you're referencing the Solr QueryElevationComponent. https://wiki.apache.org/solr/QueryElevationComponent

There has been previous discussion on github as to how to emulate this functionality in Elasticsearch. https://github.com/elasticsearch/elasticsearch/issues/1066

The idea is to use the Elasticsearch Constant Score Query to apply a constant boost for matching documents. Apply a big enough boost and you in effect get the "sponsored search" result of Solr's Elevate.

From the documentation:

Constant Score Query

A query that wraps a filter or another query and simply returns a constant score equal to the query boost for every document in the filter. Maps to Lucene ConstantScoreQuery.

{
    "constant_score" : {
        "filter" : {
            "term" : { "user" : "kimchy"}
        },
        "boost" : 1.2
    }
}

The filter object can hold only filter elements, not queries. Filters can be much faster compared to queries since they don’t perform any scoring, especially when they are cached.

A query can also be wrapped in a constant_score query:

{
    "constant_score" : {
        "query" : {
            "term" : { "user" : "kimchy"}
        },
        "boost" : 1.2
    }
}

http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/query-dsl-constant-score-query.html

EDIT

Example using multiple boosts:

{
    "query": {
        "bool": {
            "should": [{
                "constant_score": {
                    "filter": {
                        "terms": {
                            "user": ["kimchy", "john"]
                        }
                    },
                    boost: 3
                }
            }, {
                "constant_score": {
                    "filter": {
                        "terms": {
                            "user": ["paul", "ringo"]
                        }
                    },
                    boost: 2
                }
            }]
        }
    }
}

Upvotes: 2

Related Questions