Re: PostgreSQL Mailing list public archives : search not working ...

Hi,

Theory: Maybe the search doesn't always work ...
Is this a known problem?

My use case:
in /list/ -> "Search the Archives"
search for the "large database backup"
->
/search/?m=1&q=large+database+backup&l=&d=365&s=r
-> "Your search for large database backup returned no hits."

But at least one big thread exists:

To: pgsql-general <pgsql-general(at)lists(dot)postgresql(dot)org>
Subject: postgres *large database backup*
Date: 2022-11-30 15:40:23
/message-id/flat/CA%2BONtZ7CdiJ2X8x_8sPHR%2Btz-vv7mR1uoDzLia3fzWVZA9k0JQ%40mail.gmail.com

*Changing the "Sort by" ..*

- sort by: rank: no hits
/search/?m=1&q=large+database+backup&l=&d=365&s=r

- sort by: date: OK ; 1 result
/search/?m=1&q=large+database+backup&l=&d=365&s=d

- sort by: reverse date : no hits.
/search/?m=1&q=large+database+backup&l=&d=365&s=i

Regards,
Imre

[ redirecting to -www list ]

Imre Samu <pella(dot)samu(at)gmail(dot)com> writes:
> My use case:
> in /list/ -> "Search the Archives"
> search for the "large database backup"
> ->
> /search/?m=1&q=large+database+backup&l=&d=365&s=r
> -> "Your search for large database backup returned no hits."

> But at least one big thread exists:

> To: pgsql-general <pgsql-general(at)lists(dot)postgresql(dot)org>
> Subject: postgres *large database backup*
> Date: 2022-11-30 15:40:23
> /message-id/flat/CA%2BONtZ7CdiJ2X8x_8sPHR%2Btz-vv7mR1uoDzLia3fzWVZA9k0JQ%40mail.gmail.com

> *Changing the "Sort by" ..*

> - sort by: rank: no hits
> /search/?m=1&q=large+database+backup&l=&d=365&s=r

> - sort by: date: OK ; 1 result
> /search/?m=1&q=large+database+backup&l=&d=365&s=d

> - sort by: reverse date : no hits.
> /search/?m=1&q=large+database+backup&l=&d=365&s=i

I can confirm that this example gives different result sets depending
on the choice of sort order. Surely that's a bug? I have little idea
what's going on under the hood though.

regards, tom lane

>
>
> I can confirm that this example gives different result sets depending
> on the choice of sort order. Surely that's a bug? I have little idea
> what's going on under the hood though.
>

I assumed that if I searched for an existing email subject[1], I would get
a hit.
And when I didn't get a hit [2] I was very surprised and didn't understand
why?
So It's strange to me.

[1] /list/pgsql-general/since/202212010000/
"postgres large database backup" <-- so this email subject exists.

[2] if I search for "postgres large database backup" in
"pgsql-general" ordered by "date"
/search/?m=1&q=postgres+large+database+backup&l=2&d=-1&s=d
~ 2 ( + this mail expected )

compare to
https://postgrespro.com/search/?q=postgres+large+database+backup&area=archive&list=pgsql-general&period=&order=date
where I see - the expected mail
- 2022-11-30 18:40:23 | postgres large database backup (Atul Kumar)

regards,
Imre

Greetings,

* Imre Samu (pella(dot)samu(at)gmail(dot)com) wrote:
> Theory: Maybe the search doesn't always work ...

It does ... but perhaps not as expected (or as ideal).

> Is this a known problem?

Yes and no.

> My use case:
> in /list/ -> "Search the Archives"
> search for the "large database backup"
> ->
> /search/?m=1&q=large+database+backup&l=&d=365&s=r
> -> "Your search for large database backup returned no hits."

Note that the default is "within the last year". That may not be ideal
though and perhaps we should change it.

> But at least one big thread exists:

"big" is perhaps over-stating it, and most of it is actually about ZFS
anyway, though a few of the initial emails on the thread are
informative.

> To: pgsql-general <pgsql-general(at)lists(dot)postgresql(dot)org>
> Subject: postgres *large database backup*
> Date: 2022-11-30 15:40:23
> /message-id/flat/CA%2BONtZ7CdiJ2X8x_8sPHR%2Btz-vv7mR1uoDzLia3fzWVZA9k0JQ%40mail.gmail.com
>
> *Changing the "Sort by" ..*

Changing date to 'anytime', sorting by 'date' and using 'database
backup' instead turns this thread up pretty quickly (number 7):

/search/?m=1&q=database+backup&l=&d=-1&s=d

That said, we should really be including the Subject line in the search
and I don't think we are today (looks like we just search the body).
The code is available for review here:

https://git.postgresql.org/gitweb/?p=pgarchives.git;a=summary

Patches to improve certainly welcome.

Thanks,

Stephen