[
https://issues.apache.org/jira/browse/JENA-1693?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16808682#comment-16808682
]
Andy Seaborne commented on JENA-1693:
-------------------------------------
[~ajs6f] A good place to look would be the current aggregator like COUNT. If it
is a keyword in the language, it's not a custom extension.
There is support for custom aggregators, see {{AggCustom}}, it needs Java to
register code IIRC. It shoudl work but I don't think it gets used.
As a statistic, MEDIAN is not like AVG and STDDEV. They are streaming
accumulations. I don't know of an algorithm for MEDIAN that avoids a sort -
maybe there is such an algorithm that is not a full sort and retains only
partial intermediates - but retaining some values while waiting all values
seems to be necessary. So this aggregator will take memory to evaluate.
Investigation needed!
> Add Aggregate Function MEDIAN To SPARQL ARQ Syntax
> --------------------------------------------------
>
> Key: JENA-1693
> URL: https://issues.apache.org/jira/browse/JENA-1693
> Project: Apache Jena
> Issue Type: New Feature
> Environment: general
>
> Reporter: Marco Neumann
> Priority: Minor
> Labels: features
> Original Estimate: 24h
> Remaining Estimate: 24h
>
> As briefly mentioned to Andy Seaborne I'd like to see the aggregate function
> MEDIAN in the ARQ SPARQL syntax.
> "Median is the value that separates lower half from the higher half when the
> values are ordered in ascending or descending order. It is the middle value
> in a given dataset. Medians are helpful in understanding the distribution of
> data. This can be done by comparing mean and median values. By observing the
> difference between these values we can understand whether the data is left
> skewed or right skewed. The formula for median is: Median = ((n + 1)/2) th
> number in the series where the numbers are ordered. Here, n denotes the
> number of values for the given variable."
> DIVYA SPANDANA MARNEN, SPARQL-R: EXTENDED SPARQL FOR STATISTICAL COMPUTATIONS.
>
> example
>
> SELECT agg:median(?age) AS ?median
> WHERE
> { ?x ex:age ?age }
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
