# Query DSL count distinct

**URL:** https://discuss.elastic.co/t/query-dsl-count-distinct/299751
**Category:** Elasticsearch
**Created:** [March 15, 2022, 2:46pm UTC](https://discuss.elastic.co/t/query-dsl-count-distinct/299751 "2022-03-15T14:46:03Z")
**Posts on this page:** 17
**Page:** 1

<div class="post-metadata">

### Author: ![avnere](https://avatars.discourse-cdn.com/v4/letter/a/ba8739/32.png) [@avnere](https://discuss.elastic.co/u/avnere)
#### Post date: [March 15, 2022, 2:46pm UTC](https://discuss.elastic.co/t/query-dsl-count-distinct/299751/1 "2022-03-15T14:46:03Z")

</div>

Hi,

I am looking for precise count results for index with over 70 millions documents.  
Equal to this: **SQL** : select count(distinct column) from table;

I already tried Terms-aggregation, composite-aggregation, cardinality-aggregation. But didn't get exact count.

What is the best approach for it?

---

<div class="post-metadata">

### Author: ![dadoonet](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/dadoonet/32/137187_2.png) [@dadoonet](https://discuss.elastic.co/u/dadoonet)
#### Post date: [March 15, 2022, 3:03pm UTC](https://discuss.elastic.co/t/query-dsl-count-distinct/299751/2 "2022-03-15T15:03:47Z")

</div>

If you want to count the number of documents, try:

```auto
GET /your-index/_search?track_total_hits=true&size=0

```

---

<div class="post-metadata">

### Author: ![avnere](https://avatars.discourse-cdn.com/v4/letter/a/ba8739/32.png) [@avnere](https://discuss.elastic.co/u/avnere)
#### Post date: [March 15, 2022, 3:42pm UTC](https://discuss.elastic.co/t/query-dsl-count-distinct/299751/3 "2022-03-15T15:42:53Z")

</div>

I want to count distinct value of field.  
something like: select count(distinct name) from table;

---

<div class="post-metadata">

### Author: ![dadoonet](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/dadoonet/32/137187_2.png) [@dadoonet](https://discuss.elastic.co/u/dadoonet)
#### Post date: [March 15, 2022, 3:55pm UTC](https://discuss.elastic.co/t/query-dsl-count-distinct/299751/4 "2022-03-15T15:55:28Z")

</div>

So [terms aggregation](https://www.elastic.co/guide/en/elasticsearch/reference/current/search-aggregations-bucket-terms-aggregation.html) is the way to go.

---

<div class="post-metadata">

### Author: ![avnere](https://avatars.discourse-cdn.com/v4/letter/a/ba8739/32.png) [@avnere](https://discuss.elastic.co/u/avnere)
#### Post date: [March 15, 2022, 4:46pm UTC](https://discuss.elastic.co/t/query-dsl-count-distinct/299751/5 "2022-03-15T16:46:05Z")

</div>

The index with over 70 millions documents. I already tried Terms-aggregation, composite-aggregation, cardinality-aggregation. But didn't get exact count.

---

<div class="post-metadata">

### Author: ![dadoonet](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/dadoonet/32/137187_2.png) [@dadoonet](https://discuss.elastic.co/u/dadoonet)
#### Post date: [March 15, 2022, 6:03pm UTC](https://discuss.elastic.co/t/query-dsl-count-distinct/299751/6 "2022-03-15T18:03:34Z")

</div>

Could you share the request and the response please? With a `terms` agg.

---

<div class="post-metadata">

### Author: ![avnere](https://avatars.discourse-cdn.com/v4/letter/a/ba8739/32.png) [@avnere](https://discuss.elastic.co/u/avnere)
#### Post date: [March 15, 2022, 6:28pm UTC](https://discuss.elastic.co/t/query-dsl-count-distinct/299751/7 "2022-03-15T18:28:30Z")

</div>

```auto
GET abc/_search
{
"aggs":{
        "unique_fields": {
            "terms": {
                "field": "source",
                "size": 10000,
                "order": {
                  "_key": "asc"
                }
            }
        }
    },
    "size": 0
}

```

In order to compare I ran count on each "source" field and saw that the values are not equal.

---

<div class="post-metadata">

### Author: ![dadoonet](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/dadoonet/32/137187_2.png) [@dadoonet](https://discuss.elastic.co/u/dadoonet)
#### Post date: [March 15, 2022, 7:11pm UTC](https://discuss.elastic.co/t/query-dsl-count-distinct/299751/8 "2022-03-15T19:11:24Z")

</div>

Could you share the output?  
I'd like to see the first part (with some first terms)

---

<div class="post-metadata">

### Author: ![avnere](https://avatars.discourse-cdn.com/v4/letter/a/ba8739/32.png) [@avnere](https://discuss.elastic.co/u/avnere)
#### Post date: [March 15, 2022, 8:31pm UTC](https://discuss.elastic.co/t/query-dsl-count-distinct/299751/9 "2022-03-15T20:31:08Z")

</div>

```auto
{
  "took" : 2,
  "timed_out" : false,
  "_shards" : {
    "total" : 1,
    "successful" : 1,
    "skipped" : 0,
    "failed" : 0
  },
  "hits" : {
    "total" : {
      "value" : 10000,
      "relation" : "gte"
    },
    "max_score" : null,
    "hits" : []
  },
  "aggregations" : {
    "unique_fields" : {
      "doc_count_error_upper_bound" : 0,
      "sum_other_doc_count" : 0,
      "buckets" : [
        {
          "key" : "FIFO",
          "doc_count" : 13
        },
        {
          "key" : "FW",
          "doc_count" : 77757759
        },
        {
          "key" : "HW",
          "doc_count" : 12394396
        },
        {
          "key" : "Info",
          "doc_count" : 6705
        }
      ]
    }
  }
}

```

---

<div class="post-metadata">

### Author: ![dadoonet](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/dadoonet/32/137187_2.png) [@dadoonet](https://discuss.elastic.co/u/dadoonet)
#### Post date: [March 15, 2022, 10:48pm UTC](https://discuss.elastic.co/t/query-dsl-count-distinct/299751/10 "2022-03-15T22:48:30Z")

</div>

The numbers you are seeing are the exact numbers. What makes you think that they are not?

---

<div class="post-metadata">

### Author: ![casterQ](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/casterq/32/93257_2.png) [@casterQ](https://discuss.elastic.co/u/casterQ)
#### Post date: [March 16, 2022, 2:30am UTC](https://discuss.elastic.co/t/query-dsl-count-distinct/299751/11 "2022-03-16T02:30:58Z")

</div>

terms agg, you can use `shard_size`

> **[Terms aggregation | Elasticsearch Guide \[8.1\] | Elastic](https://www.elastic.co/guide/en/elasticsearch/reference/8.1/search-aggregations-bucket-terms-aggregation.html#search-aggregations-bucket-terms-aggregation-shard-size)**

cardinality agg, you can use precision\_threshold  
[![image](https://us1.discourse-cdn.com/elastic/original/3X/9/f/9f574e0fa4fa751a50efc38dddc3818ab0d0160d.png)](https://www.elastic.co/guide/en/elasticsearch/reference/8.1/search-aggregations-bucket-terms-aggregation.html#search-aggregations-bucket-terms-aggregation-shard-size)

---

<div class="post-metadata">

### Author: ![avnere](https://avatars.discourse-cdn.com/v4/letter/a/ba8739/32.png) [@avnere](https://discuss.elastic.co/u/avnere)
#### Post date: [March 16, 2022, 7:48am UTC](https://discuss.elastic.co/t/query-dsl-count-distinct/299751/12 "2022-03-16T07:48:18Z")

</div>

When running count on specific field I get different number.

For Example:  
When running the following I get count 6561, while with terms aggs I got 6705.

```auto
GET baseelements_rwr/_count
{
  "query": {
    "bool": {
        "must": [
            {"match": {"source": "Info"}}
        ]
    }
  }
}

```

---

<div class="post-metadata">

### Author: ![Christian\_Dahlqvist](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/christian_dahlqvist/32/4617_2.png) [@Christian\_Dahlqvist](https://discuss.elastic.co/u/Christian_Dahlqvist)
#### Post date: [March 16, 2022, 7:58am UTC](https://discuss.elastic.co/t/query-dsl-count-distinct/299751/13 "2022-03-16T07:58:03Z")

</div>

Cardinality aggregation, like some other aggregation types, are by design approximations. This is generally highligted in the documentation. I recommend reading the explanation and discussion provided [in this thread](https://discuss.elastic.co/t/elasticsearch-aggregation-on-million-or-more-data/293641) as well as [this one](https://discuss.elastic.co/t/are-there-plans-to-support-accurate-distinct/232808).

---

<div class="post-metadata">

### Author: ![dadoonet](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/dadoonet/32/137187_2.png) [@dadoonet](https://discuss.elastic.co/u/dadoonet)
#### Post date: [March 16, 2022, 9:48am UTC](https://discuss.elastic.co/t/query-dsl-count-distinct/299751/14 "2022-03-16T09:48:02Z")

</div>

Could you try with:

```auto
GET baseelements_rwr/_count
{
  "query": {
    "bool": {
        "must": [
            {"term": {"source": "Info"}}
        ]
    }
  }
}

```

What is the mapping for this `source` field?  
What does a typical document look like?

---

<div class="post-metadata">

### Author: ![avnere](https://avatars.discourse-cdn.com/v4/letter/a/ba8739/32.png) [@avnere](https://discuss.elastic.co/u/avnere)
#### Post date: [March 16, 2022, 10:34am UTC](https://discuss.elastic.co/t/query-dsl-count-distinct/299751/15 "2022-03-16T10:34:37Z")

</div>

source field is keyword.

Same results when switching to term instead match.

---

<div class="post-metadata">

### Author: ![dadoonet](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/dadoonet/32/137187_2.png) [@dadoonet](https://discuss.elastic.co/u/dadoonet)
#### Post date: [March 16, 2022, 11:09am UTC](https://discuss.elastic.co/t/query-dsl-count-distinct/299751/16 "2022-03-16T11:09:35Z")

</div>

Could you share a typical document please?

---

<div class="post-metadata">

### Author: ![system](https://us1.discourse-cdn.com/elastic/original/3X/1/a/1ac57faf039f6b580b3f104ef42a2a89e41014de.png) [@system](https://discuss.elastic.co/u/system)
#### Post date: [April 13, 2022, 11:09am UTC](https://discuss.elastic.co/t/query-dsl-count-distinct/299751/17 "2022-04-13T11:09:56Z")

</div>

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