Hi,
If you call the action of "sifting" ordering, then yes. If you don't call
it ordering, then no.

In essence, is the output of a filtering mechanism, done in a single result
set pass. And this pass should be the same pass in charge of collecting the
result set in the first place.

Thanks


On Mon, May 13, 2024 at 5:48 AM Wolfgang Wilhelm <wolfgang20121...@yahoo.de>
wrote:

> Hi,
>
> do I interpret your idea correctly: You want some sort of ordering without
> ordering?
>
> Kind regards
> WW
>
> Am Montag, 13. Mai 2024 um 10:40:38 MESZ hat aa <ghe...@gmail.com>
> Folgendes geschrieben:
>
>
> Hello Everyone!
>
> Is there any chance to get some kind of a result set sifting mechanism in
> Postgres?
>
> What I am looking for is a way to get for example: "nulls last" in a
> result set, without having to call "order by" or having to use UNION ALL,
> and if possible to get this in a single result set pass.
>
> Something on this line: SELECT a, b, c FROM my_table WHERE a nulls last
> OFFSET 0 LIMIT 25
>
> I don't want to use order by or union all because these are time consuming
> operations, especially on  large data sets and when comparations are done
> on dynamic values (eg: geolocation distances in between a mobile and a
> static location)
>
> What I would expect from such a feature, will be speeds comparable with
> non sorted selects, while getting a very rudimentary ordering.
>
> A use case for such a mechanism will be the implementation of QUICK
> relevant search results for a search engine.
>
> I'm not familiar with how Postgres logic handles simple select queries,
> but the way I would envision a result set sifting logic, would be to
> collect the result set, in 2 separate lists, based on the sifting
> condition, and then concatenate these 2 lists and return the result, when
> the pagination requests conditions are met.
>
> Any idea if such a functionality is feasible ?
>
> Thank you.
>
>   PS: if ever implemented, the  sifting mechanism could be extended to
> accommodate any type of thresholds, not just null values.
>
>
>
>

Reply via email to