Message-ID: <e665a0c8-4a5c-4345-2234-4487138a0d6f@enterprisedb.com>
Date: Thu, 28 Dec 2023 15:24:59 +0100
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.15.1
Subject: Re: Multidimensional Histograms
Content-Language: en-US
From: Tomas Vondra <tomas.vondra@enterprisedb.com>
To: Alexander Cheshev <alex.cheshev@gmail.com>, pgsql-hackers@postgresql.org
References: <CAN_hQmsYDhPMK4bSJHPw3RGnpKqPbx2Uk_7HRV+Dz9F5H7dj1A@mail.gmail.com>
 <6e92450c-8136-11d4-8f6e-501c693af5c8@enterprisedb.com>
In-Reply-To: <6e92450c-8136-11d4-8f6e-501c693af5c8@enterprisedb.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Archived-At: <https://www.postgresql.org/message-id/e665a0c8-4a5c-4345-2234-4487138a0d6f%40enterprisedb.com>
Precedence: bulk

On 12/27/23 22:19, Tomas Vondra wrote:
> Hello Alexander,
> 
> We actually had histograms in the early patches adding multivariate
> statistics [1]. We however ended up removing histograms and only kept
> the simpler types, for a couple reasons.
> 
> It might be worth going through the discussion - I'm sure one of the
> reasons was we needed to limit the scope of the patch, and histograms
> were much more complex and possibly less useful compared to the other
> statistics types.
>
> ...

FWIW I did not intend to reject the idea of adding multi-dimensional
histograms, but rather to provide some insight into the history of the
past attempt, and also point some weaknesses of the algorithm described
in the 1988 paper. If you choose to work on this, I can do a review etc.

regards

-- 
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company