Mailing-List: contact dev-help@hive.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@hive.apache.org
Date: Mon, 20 Mar 2017 13:17:41 +0000 (UTC)
From: "Carter Shanklin (JIRA)" <jira@apache.org>
To: dev@hive.apache.org
Message-ID: <JIRA.13057521.1490015815000.65047.1490015861597@Atlassian.JIRA>
In-Reply-To: <JIRA.13057521.1490015815000@Atlassian.JIRA>
References: <JIRA.13057521.1490015815000@Atlassian.JIRA> <JIRA.13057521.1490015815148@jira-lw-us.apache.org>
Subject: [jira] [Created] (HIVE-16255) Support percentile_cont /
 percentile_disc
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
archived-at: Mon, 20 Mar 2017 13:17:45 -0000

Carter Shanklin created HIVE-16255:
--------------------------------------

             Summary: Support percentile_cont / percentile_disc
                 Key: HIVE-16255
                 URL: https://issues.apache.org/jira/browse/HIVE-16255
             Project: Hive
          Issue Type: Sub-task
            Reporter: Carter Shanklin


Way back in HIVE-259, a percentile function was added that provides a subse=
t of the standard percentile_cont aggregate function.

The SQL standard provides some additional options and also a percentile_dis=
c aggregate function with different rules. In the standard you specify an o=
rdering with arbitrary value expression and the results are drawn from this=
 value expression. This aggregate functions should be usable as analytic fu=
nctions as well (i.e. support the over clause). The current percentile func=
tion is able to be used with an over clause.

The rough outline of how this works is:
percentile_cont(number) within group (order by expression) [ over(window sp=
ec) ]
percentile_disc(number) within group (order by expression) [ over(window sp=
ec) ]

The value of number should be between 0 and 1. The value expression is eval=
uated for each row of the group, nulls are discarded, and the remaining row=
s are ordered.

=E2=80=94 If PERCENTILE_CONT is specified, by considering the pair of conse=
cutive rows that are indicated by the argument, treated as a fraction of th=
e total number of rows in the group, and interpolating the value of the val=
ue expression evaluated for these rows.

=E2=80=94 If PERCENTILE_DISC is specified, by treating the group as a windo=
w partition of the CUME_DIST window function, using the specified ordering =
of the value expression as the window ordering, and returning the  first va=
lue expression whose cumulative distribution value is greater than or equal=
 to the argument.


--
This message was sent by Atlassian JIRA
(v6.3.15#6346)