# Elasticsearch cardinality aggregation not returning accurate numbers despite using precision\_threshold

**URL:** <https://discuss.elastic.co/t/elasticsearch-cardinality-aggregation-not-returning-accurate-numbers-despite-using-precision-threshold/288632>\
**Category:** Elasticsearch\
**Created:** [November 8, 2021, 1:50pm UTC](https://discuss.elastic.co/t/elasticsearch-cardinality-aggregation-not-returning-accurate-numbers-despite-using-precision-threshold/288632 "2021-11-08T13:50:23Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![preetish\_P](https://avatars.discourse-cdn.com/v4/letter/p/77aa72/32.png) [@preetish\_P](https://discuss.elastic.co/u/preetish_P)\
**Post date:** [November 8, 2021, 1:50pm UTC](https://discuss.elastic.co/t/elasticsearch-cardinality-aggregation-not-returning-accurate-numbers-despite-using-precision-threshold/288632/1 "2021-11-08T13:50:23Z")

</div>

Hi folks,

I am trying to show unique count in a metric and observed that the numbers are wrong.

### Setup:

I ​used enhanced table visualisation and did a "unique count" of keyword type field. in advanced setting I used `{"precision_threshold":20000}`.

The filter used on top is of the format: field1.keyword:"value" AND field2:9\*. I am getting a unique count of `13680`.

 ![Screen Shot 2021-11-08 at 13.40.35](https://us1.discourse-cdn.com/elastic/original/3X/d/6/d67a8babd697408e3477d58f1e49ce8852ad2dc8.png)

However when I do terms aggregation and calculate the sum, it is showing `13678`. I verified this number by downloading the csv and counting within excel.

 ![Screen Shot 2021-11-08 at 13.41.05](https://us1.discourse-cdn.com/elastic/original/3X/5/0/50a00e62e55381f4d2bcbbeaa600a6c9190c5dca.png)

The same issue occurs for metric type visualisation. Is there any setting which I am missing here or is this a product bug ?

Thank you

---

<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:** [November 8, 2021, 3:19pm UTC](https://discuss.elastic.co/t/elasticsearch-cardinality-aggregation-not-returning-accurate-numbers-despite-using-precision-threshold/288632/2 "2021-11-08T15:19:58Z")

</div>

Have a look at [this thread](https://discuss.elastic.co/t/cardinality-agg-off-by-one-even-after-precision-increase/283096/2).

---

<div class="post-metadata">

**Author:** ![preetish\_P](https://avatars.discourse-cdn.com/v4/letter/p/77aa72/32.png) [@preetish\_P](https://discuss.elastic.co/u/preetish_P)\
**Post date:** [November 8, 2021, 4:56pm UTC](https://discuss.elastic.co/t/elasticsearch-cardinality-aggregation-not-returning-accurate-numbers-despite-using-precision-threshold/288632/3 "2021-11-08T16:56:34Z")

</div>

Hi @Christian_Dahlqvist ,

In my case its is a single index, hence the field is consistently of the type `keyword`.  
Does it mean it is a limitation of the 1st counting technique ?

Thanks

---

<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:** [November 8, 2021, 5:10pm UTC](https://discuss.elastic.co/t/elasticsearch-cardinality-aggregation-not-returning-accurate-numbers-despite-using-precision-threshold/288632/4 "2021-11-08T17:10:15Z")

</div>

As far as I know it is always an approximation.

---

<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:** [December 6, 2021, 5:10pm UTC](https://discuss.elastic.co/t/elasticsearch-cardinality-aggregation-not-returning-accurate-numbers-despite-using-precision-threshold/288632/5 "2021-12-06T17:10:55Z")

</div>

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