Getting unique results with elasticsearch acorrding to field - php

Question

I'm using FOSElasticaBundle with Symfony2 on my project and there are entry and user tables on MySQL database and each entry belongs to one user.
I want to get just one entry per a user among the whole entries from the database.
Entries Representation
[
{
"id": 1,
"name": "Hello world",
"user": {
"id": 17,
"username": "foo"
}
},
{
"id": 2,
"name": "Lorem ipsum",
"user": {
"id": 15,
"username": "bar"
}
},
{
"id": 3,
"name": "Dolar sit amet",
"user": {
"id": 17,
"username": "foo"
}
},
]
Expected result is:
[
{
"id": 1,
"name": "Hello world",
"user": {
"id": 17,
"username": "foo"
}
},
{
"id": 2,
"name": "Lorem ipsum",
"user": {
"id": 15,
"username": "bar"
}
}
]
But it returns all entries on table. I've tried to add an aggregation to my elasticsearch query and nothing changed.
$distinctAgg = new \Elastica\Aggregation\Terms("distinctAgg");
$distinctAgg->setField("user.id");
$distinctAgg->setSize(1);
$query->addAggregation($distinctAgg);
Is there any way to do this via term filter or anything else? Any help would be great. Thank you.

Answer 1

Aggregations are not easy to understand when you are used to MySQL group by.
The first thing, is that aggregations results are not returned in hits, but in aggregations. So when you get the result of your search, you have to get aggregations like that :
$results = $search->search();
$aggregationsResults = $results->getAggregations();
The second thing is that aggregations wont return you the source. With the aggregation of your example, you will only know that you have 1 user with ID 15, and 2 users with ID 15.
E.g. with this query :
{
"query": {
"match_all": {}
},
"aggs": {
"byUser": {
"terms": {
"field": "user.id"
}
}
}
}
Result:
{
"took": 1,
"timed_out": false,
"_shards": {
"total": 5,
"successful": 5,
"failed": 0
},
"hits": {
"total": 3,
"max_score": 1,
"hits": [ ... ]
},
"aggregations": {
"byUser": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [
{
"key": 17,
"doc_count": 2
},
{
"key": 15,
"doc_count": 1
}
]
}
}
}
If you want to get results, the same way you would do with a GROUP BY in MySQL, you have to use a top_hits sub-aggregation:
{
"query": {
"match_all": {}
},
"aggs": {
"byUser": {
"terms": {
"field": "user.id"
},
"aggs": {
"results": {
"top_hits": {
"size": 1
}
}
}
}
}
}
Result:
{
"took": 3,
"timed_out": false,
"_shards": {
"total": 5,
"successful": 5,
"failed": 0
},
"hits": {
"total": 3,
"max_score": 1,
"hits": [ ... ]
},
"aggregations": {
"byUser": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [
{
"key": 17,
"doc_count": 2,
"results": {
"hits": {
"total": 2,
"max_score": 1,
"hits": [
{
"_index": "test_stackoverflow",
"_type": "test1",
"_id": "1",
"_score": 1,
"_source": {
"id": 1,
"name": "Hello world",
"user": {
"id": 17,
"username": "foo"
}
}
}
]
}
}
},
{
"key": 15,
"doc_count": 1,
"results": {
"hits": {
"total": 1,
"max_score": 1,
"hits": [
{
"_index": "test_stackoverflow",
"_type": "test1",
"_id": "2",
"_score": 1,
"_source": {
"id": 2,
"name": "Lorem ipsum",
"user": {
"id": 15,
"username": "bar"
}
}
}
]
}
}
}
]
}
}
}
More informations on this page : https://www.elastic.co/blog/top-hits-aggregation