Date: 2020-01-10 09:07:03 +0100
From: @PedroTadim
To: SQL devs <>
Version: 11.35.9 (Nov2019-SP1)

Last updated: 2020-02-21 12:36:58 +0100

Comment 27508

Date: 2020-01-10 09:07:03 +0100
From: @PedroTadim

User-Agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/79.0.3945.117 Safari/537.36
Build Identifier:

If median_avg or the quantile_avg of a group is going to be computed as the average of two values, if one of them is NULL, then the value is ignored, while it shouldn't.

Reproducible: Always

Steps to Reproduce:

1. SELECT median_avg(c1) FROM (VALUES (1), (NULL)) t1(c1);

Actual Results:

-128

Expected Results:

NULL

Comment 27509

Date: 2020-01-10 09:50:00 +0100
From: MonetDB Mercurial Repository <>

Changeset 14395b0ea40e made by Pedro Ferreira pedro.ferreira@monetdbsolutions.com in the MonetDB repo, refers to this bug.

For complete details, see https//devmonetdborg/hg/MonetDB?cmd=changeset;node=14395b0ea40e

Changeset description:

Added test and fix for bug #6807, ie look for NULL values while computing quantile_avg

Comment 27510

Date: 2020-01-10 11:33:12 +0100
From: @sjoerdmullender

Is this the correct thing to do?
Shouldn't NULL be ignored when calculating the median/quantile?
This would mean that the result is only NULL if all values are NULL, not if any is NULL.