Re: BUG #13619: regression functions return Null - Mailing list pgsql-bugs

From David G. Johnston
Subject Re: BUG #13619: regression functions return Null
Date
Msg-id CAKFQuwaeo0o0e4UeHEbVgbv-iOwgsaj=aaXAHpKQc6vgZ_e_sQ@mail.gmail.com
Whole thread Raw
In response to BUG #13619: regression functions return Null  (biswadeep.banerjee@actifio.com)
List pgsql-bugs
On Mon, Sep 14, 2015 at 6:14 AM, <biswadeep.banerjee@actifio.com> wrote:

> The following bug has been logged on the website:
>
> Bug reference:      13619
> Logged by:          Biswadeep Banerjee
> Email address:      biswadeep.banerjee@actifio.com
> PostgreSQL version: 9.4.4
> Operating system:   Any
> Description:
>
> Hi,
> Trending analysis is one of our use case, where I am trying the use the
> regression function in postgres 9.4.4.
> eg. regr_slope(y,x); regr_intercept(y,x).
>
> I have tried to use the functions to fit in our trending analysis SQL but
> returns NULL.
> As test also tried with constant values, like,
> select regr_slope(4,3), regr_intercept(4,3) from table;
>
> --- returns NULL.
>
> Is this a known issue? or I am using the function differently.
>

=E2=80=8BThese are "aggregate" functions and you need to use them in a quer=
y that
contains a "GROUP BY".

Think about it this way - what is the slope and intercept of a single
point?  That is basically what you are asking when you write
"regr_slope(4,3)"

David J.
=E2=80=8B

pgsql-bugs by date:

Previous
From: biswadeep.banerjee@actifio.com
Date:
Subject: BUG #13619: regression functions return Null
Next
From: "David G. Johnston"
Date:
Subject: Re: BUG #13619: regression functions return Null