I think the subquery approach should be something like:
---
select id, pg_sleep(0.1) from (select id from thing offset 90 order by tag)
last_things
---
Is that right?
Then what if the interface is exposed as a view? e.g.,
---
create view thing_interface as select id, tag, pg_sleep(0.1) from
On Sunday, December 6, 2020, 孙冰 wrote:
> The skipped rows by an OFFSET clause have to be computed nevertheless. I
> am wondering if there could be any chance to improve, since the computation
> is on the *entire* rows rather than on the *criterial* columns.
>
> [...]
>
> I don't understand the
The skipped rows by an OFFSET clause have to be computed nevertheless. I am
wondering if there could be any chance to improve, since the computation is
on the *entire* rows rather than on the *criterial* columns.
Consider the following example.
Create a sample table and insert some random data.