You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Gengliang Wang (Jira)" <ji...@apache.org> on 2026/01/16 16:42:00 UTC

[jira] [Commented] (SPARK-53886) Percentile estimation functions

    [ https://issues.apache.org/jira/browse/SPARK-53886?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18052447#comment-18052447 ] 

Gengliang Wang commented on SPARK-53886:
----------------------------------------

[~cboumalh] quick check: any updates on these new functions?

> Percentile estimation functions
> -------------------------------
>
>                 Key: SPARK-53886
>                 URL: https://issues.apache.org/jira/browse/SPARK-53886
>             Project: Spark
>          Issue Type: New Feature
>          Components: SQL
>    Affects Versions: 0.5.0
>            Reporter: Gengliang Wang
>            Priority: Major
>
> Similar to https://issues.apache.org/jira/browse/SPARK-53885, it would be useful to have the following percentile estimation functions based on data sketch. These functions provide a lightweight, mergeable representation of percentile distributions, enabling efficient approximate percentile computation over large or streaming datasets without maintaining full data samples.
>  * *APPROX_PERCENTILE_ACCUMULATE(expr[, maxItemsTracked]):* Creates an intermediate state object that incrementally accumulates values for percentile estimation using a sketch-based algorithm.
>  * *APPROX_PERCENTILE_ESTIMATE(state [, k] ])* : Returns the approximate percentile value(s) from a previously accumulated sketch state.
>  * *APPROX_PERCENTILE_COMBINE(expr[, maxItemsTracked])* : Merges two intermediate sketch states to enable distributed or parallel percentile estimation.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org