> For the complete documentation index, see [llms.txt](https://docs.pinot.apache.org/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.pinot.apache.org/release-1.0.0/configuration-reference/functions/percentilekll.md).

# percentilekll

This section contains reference documentation for the PERCENTILEKLL function.

`KLL Sketch` is an approxiamate quantiles algorithm which targets optimal space for a given accuracy. `PERCENTILEKLL` is a percentile calculation aggregation function based on Apache Datasketches [KLL Doubles Sketch](https://datasketches.apache.org/docs/KLL/KLLSketch.html) implementation.

Pinot also offers a 'raw' variant, `PERCENTILEKLLRAW`, which returns the serialized sketch that can be used for calculating 'rank' or 'histogram'.

All of the variants of `PercentileKLL` also support raw sketches in Pinot columns. This means you can create KLL Doubles sketches outside of Pinot and ingest them into columns as binary strings. `PercentileKLL` will identify these columns merge them to produce aggregate results.

## Signature

> PercentileKLL(column, percentile, kValue) -> Double

* `column` (required): Name of the column to aggregate on. If the column is a multi value column, use `PERCENTILEKLLMV` variant.
* `percentile` (required): Percentile value to be calculated \[0..100]
* `kValue`: Integer value which determines the size of the sketch. Default value is `200` which corresponds to a normalized rank error of about 1.65%. For details please see the [accuracy vs size chart](https://datasketches.apache.org/docs/KLL/KLLAccuracyAndSize.html).

## Usage Examples

```sql
select percentileKLL(ArrDelayMinutes, 90) as DelayP90
from airlineStats
```

| DelayP90 |
| -------- |
| 40       |

```sql
select Carrier, percentileKll(ArrDelay, 50, 600) as MedianDelay
from airlineStats
where ArrDelay > 0
group by Carrier
order by 2 desc
limit 3
```

| Carrier | MedianDelay |
| ------- | ----------- |
| MQ      | 28          |
| B6      | 28          |
| EV      | 24          |


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.pinot.apache.org/release-1.0.0/configuration-reference/functions/percentilekll.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.