____;
ERROR: channel name too long
UNLISTEN notify_async_channel_name_too_long_%._;
ERROR: channel name too long
Regards,
Alexander Cheshev
On Sun, 21 Jul 2024 at 21:36, Emanuel Calvo <3man...@gmail.com> wrote:
>
> Hi Alexander,
>
> I did
email I forgot to mention that the new implementation of
the function Exec_UnlistenCommit has the same space and time complexities
as the original implementation (which doesn't support wildcards).
Regards,
Alexander Cheshev
On Sat, 13 Jul 2024 at 13:26, Alexander Cheshev
wrote:
> Hi
[1] https://www.postgresql.org/message-id/flat/52693FC5.7070507%40gmail.com
Regards,
Alexander Cheshev
Regards,
Alexander Cheshev
On Tue, 9 Jul 2024 at 11:01, Emanuel Calvo <3man...@gmail.com> wrote:
>
> Hello there,
>
>
> El vie, 15 mar 2024 a las 9:01, Alexander Cheshev ()
> escrib
So, no overhead at all!
Regards,
Alexander Cheshev
On Mon, 8 Jan 2024 at 04:12, Andrei Lepikhov wrote:
>
> On 8/1/2024 01:36, Tomas Vondra wrote:
> > On 1/7/24 18:26, Andrei Lepikhov wrote:
> >> On 7/1/2024 17:51, Tomas Vondra wrote:
> >>> On 1/7/24 11:22, Andrei Lepik
nsional case is an NP-hard problem.
Regards,
Alexander Cheshev
On Mon, 8 Jan 2024 at 00:29, Tomas Vondra wrote:
>
>
>
> On 1/7/24 23:53, Alexander Cheshev wrote:
> > Hi Tomas,
> >
> >> The thing I was proposing is that it should be possible to build
> >&g
ion time and accuracy?
Time complexity of Equi-Depth Histogram has no dependence on B.
> But all of this can apply to histograms in general, no? It's not somehow
> special to equi-depth histograms, a v-optimal histogram could be stored
> as an r-tree etc.
I agree.
Regards,
Alexande
Hi Tomas,
I am sorry I didn't look into the code carefully. Indeed Postgres uses
Equi-Depth Histogram:
delta = (nvals - 1) / (num_hist - 1);
Regards,
Alexander Cheshev
On Sat, 6 Jan 2024 at 01:00, Alexander Cheshev wrote:
>
> Hi Tomas,
>
> > Another reason was that the al
Depth Histogram but only in 1-dimensional case.
* I will show experimental results which will demonstrate improvement
of selectivity estimation.
Then will the path be accepted by the community?
If the above path is accepted by the community then I will proceed
further with M-dimensional Equi-Depth Histog
only useful for columns with small number of distinct values;
* functional dependencies only reflect dependencies among columns
(not column values).
[1] http://www.cs.cmu.edu/~rcarlson/docs/RyanCarlson_databases.pdf
[2] https://dl.acm.org/doi/pdf/10.1145/50202.50205
--
Regards,
Alexander Cheshev
