Thread: Todo item: distinct clause - guidance

Hi,

I was looking for some to-do item to pick up and came across 'Implement DISTINCT clause in window aggregates'. Can someone please elaborate what is expectation here? Do feature request still exists as DISTINCT worked as expected if we try functions like avg(distinct col)?

Thanks,

Ankit

On Sun, 2022-11-06 at 18:39 +0530, Ankit Kumar Pandey wrote:
> I was looking for some to-do item to pick up and came across 'Implement DISTINCT
> clause in window aggregates'. Can someone pleaseelaborate what is expectation here?
> Do feature request still exists as DISTINCT worked as expected if we try functions
> like avg(distinct col)?

Whis is about window functions, like

  WITH q(a,b) AS (VALUES (1, 1), (1, 1), (2, 1))
  SELECT *,
         count(DISTINCT a) OVER (PARTITION BY b)
  FROM q;

  ERROR:  DISTINCT is not implemented for window functions
  LINE 3:        count(DISTINCT a) OVER (PARTITION BY b)
                 ^

Yours,
Laurenz Albe

On 07/11/22 00:42, Laurenz Albe wrote:
> On Sun, 2022-11-06 at 18:39 +0530, Ankit Kumar Pandey wrote:
>> I was looking for some to-do item to pick up and came across 'Implement DISTINCT
>> clause in window aggregates'. Can someone pleaseelaborate what is expectation here?
>> Do feature request still exists as DISTINCT worked as expected if we try functions
>> like avg(distinct col)?
> Whis is about window functions, like
>
>    WITH q(a,b) AS (VALUES (1, 1), (1, 1), (2, 1))
>    SELECT *,
>           count(DISTINCT a) OVER (PARTITION BY b)
>    FROM q;
>
>    ERROR:  DISTINCT is not implemented for window functions
>    LINE 3:        count(DISTINCT a) OVER (PARTITION BY b)
>                   ^
>
> Yours,
> Laurenz Albe

Thanks a lot Laurenz, I got the issue now. I was looking at aggregation 
function without windowing.

I will revisit this again.


Thanks,

Ankit