Subject: Re: list of extended statistics on psql
To: Julien Rouhaud <rjuju123@gmail.com>,
 Tatsuro Yamada <tatsuro.yamada.tf@nttcom.co.jp>
Cc: Alvaro Herrera <alvherre@2ndquadrant.com>,
 Tomas Vondra <tomas.vondra@2ndquadrant.com>,
 Michael Paquier <michael@paquier.xyz>,
 Pavel Stehule <pavel.stehule@gmail.com>,
 PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
References: <20210107144713.GA8872@alvherre.pgsql>
 <61cba5af-5a58-6bda-610c-420077b780c3@enterprisedb.com>
 <1045d2cb-7ecd-e6bf-cb50-8eeea89cc4b0@nttcom.co.jp_1>
 <9e76ce7a-e7ab-5f08-9938-59fffc20d106@enterprisedb.com>
 <29e16ceb-8777-4297-19c8-245242dd7473@enterprisedb.com>
 <876efa7e-2067-a4f9-d7e8-f6025440c870@nttcom.co.jp_1>
 <0e84aa72-25f2-1849-753f-f827949132b3@enterprisedb.com>
 <85e7e1e7-c2f0-a4a0-61de-6e7ecaee44e3@nttcom.co.jp_1>
 <79d3a4c5-1efe-5b9e-c06e-3a907b531013@nttcom.co.jp_1>
 <20210115084741.GA91293@nol>
 <2eb51458-7475-17d3-cb13-b3c950b2f930@enterprisedb.com>
From: Tomas Vondra <tomas.vondra@enterprisedb.com>
Message-ID: <5e0fb573-3c38-ee39-885c-bdc01f116335@enterprisedb.com>
Date: Sun, 17 Jan 2021 00:32:10 +0100
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101
 Thunderbird/78.6.0
MIME-Version: 1.0
In-Reply-To: <2eb51458-7475-17d3-cb13-b3c950b2f930@enterprisedb.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Precedence: bulk


On 1/15/21 5:19 PM, Tomas Vondra wrote:
> 
> 
> On 1/15/21 9:47 AM, Julien Rouhaud wrote:
>> On Wed, Jan 13, 2021 at 10:22:05AM +0900, Tatsuro Yamada wrote:
>>> Hi Tomas,
>>>
>>> On 2021/01/13 7:48, Tatsuro Yamada wrote:
>>>> On 2021/01/12 20:08, Tomas Vondra wrote:
>>>>> On 1/12/21 2:57 AM, Tatsuro Yamada wrote:
>>>>>> On 2021/01/09 9:01, Tomas Vondra wrote:
>>>>> ...>
>>>>>>> While working on that, I realized that 'defined' might be a bit
>>>>>>> ambiguous, I initially thought it means 'NOT NULL' (which it does not).
>>>>>>> I propose to change it to 'requested' instead. Tatsuro, do you agree, or
>>>>>>> do you think 'defined' is better?
>>>>>>
>>>>>> Regarding the status of extended stats, I think the followings:
>>>>>>
>>>>>>    - "defined": it shows the extended stats defined only. We can't know
>>>>>>                 whether it needs to analyze or not. I agree this name was
>>>>>>                  ambiguous. Therefore we should replace it with a more suitable
>>>>>>                 name.
>>>>>>    - "requested": it shows the extended stats needs something. Of course,
>>>>>>                 we know it needs to ANALYZE because we can create the patch.
>>>>>>                 However, I feel there is a little ambiguity for DBA.
>>>>>>                 To solve this, it would be better to write an explanation of
>>>>>>                 the status in the document. For example,
>>>>>>
>>>>>> ======
>>>>>> The column of the kind of extended stats (e. g. Ndistinct) shows some statuses.
>>>>>> "requested" means that it needs to gather data by ANALYZE. "built" means ANALYZE
>>>>>>    was finished, and the planner can use it. NULL means that it doesn't exists.
>>>>>> ======
>>>>>>
>>>>>> What do you think? :-D
>>>>>>
>>>>>
>>>>> Yes, that seems reasonable to me. Will you provide an updated patch?
>>>>
>>>>
>>>> Sounds good. I'll send the updated patch today.
>>>
>>>
>>>
>>> I updated the patch to add the explanation of the extended stats' statuses.
>>> Please feel free to modify the patch to improve it more clearly.
>>>
>>> The attached files are:
>>>     0001: Add psql \dx and the fixed document
>>>     0002: Regression test for psql \dX
>>>     app-psql.html: Created by "make html" command (You can check the
>>>                    explanation of the statuses easily, probably)
>>
>> Hello Yamada-san,
>>
>> I reviewed the patch and don't have specific complaints, it all looks good!
>>
>> I'm however thinking about the "requested" status.  I'm wondering if it could
>> lead to people think that an ANALYZE is scheduled and will happen soon.
>> Maybe "defined" or "declared" might be less misleading, or even "waiting for
>> analyze"?
>>
> 
> Well, the "defined" option is not great either, because it can be
> interpreted as "NOT NULL" - that's why I proposed "requested". Not sure
> about "declared" - I wouldn't use it in this context, but I'm not a
> native speaker so maybe it's OK.
> 

I've pushed this, keeping the "requested". If we decide that some other 
term is a better choice, we can tweak that later of course.

Thanks Tatsuro-san for the patience!


regards

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