How do I group documents in elasticsearch by a sin

Question

If I have a bunch of documents in elaticsearch that I want to be returned grouped by the one field of the document, how do I do it? Also I need it to consistently return a fixed number of results (using set maxresults)

For example if I have a bunch of documents each document representing a person and fields of the document containing attributes of the person. Let's say each person has a city field in the document. I would like to query Elasticsearch in a way that will return 50 results that are grouped by city. By 50 results I want to know how it is possible to return 50 cities mapped to all the people in those cities.

I found an implementation in:

http://www.elastic.co/guide/en/elasticsearch/reference/current/search-aggregations-metrics-top-hits-aggregation.html

But i want to apply pagination to these results as well. I dont see a setOffset and setLimit possibility in ES. Ideas?

Answer 1

how it is possible to return 50 cities mapped to all the people in those cities.

Query you are looking for looks like this:

curl -XGET 'http://localhost:9200/users/user/_search?pretty' -d '{
    "aggs": {
        "users-by-city": {
            "terms": {
                "field": "city",
                "size": 50
            },
            "aggs": {
                "top_tag_hits": {
                    "top_hits": {
                        "from": 0,
                        "size": 9000
                    }
                }
            }
        }
    }
}'

In Elastica, equivalent query could be created this way:

$query = new Elastica\Query();
$qb = new Elastica\QueryBuilder();

$query->addAggregation(
    $qb->aggregation()->terms('users-by-city')
        ->setField('city')
        ->setSize(50)
        ->addAggregation(
            $qb->aggregation()->top_hits('top-hits-in-city')
                ->setFrom(0)
                ->setSize(9000)
        )
);

If you want to paginate results, just change arguments passed into setFrom and setSize.