Feedback-ID: i76984098:Fastmail
Message-ID: <089015d7-ba30-4113-a29f-248caca375f7@aklaver.com>
Date: Thu, 23 May 2024 09:15:27 -0700
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: Json table/column design question
To: Skorpeo Skorpeo <skorpeo11@gmail.com>, pgsql-general@lists.postgresql.org
References: <CAMXrOMOav1m8Ep01o7_GDLWxLSpgMp+o4d3FNdmOY7RoWe5B7Q@mail.gmail.com>
 <f053e31c3c2e40541406ce4c266818dbbcfca7b2.camel@cybertec.at>
 <CAMXrOMOY_ky6z8W4Fh=FoWdPB5xd12ZBeCj4JMqiu4eU7s97hQ@mail.gmail.com>
Content-Language: en-US
From: Adrian Klaver <adrian.klaver@aklaver.com>
In-Reply-To: <CAMXrOMOY_ky6z8W4Fh=FoWdPB5xd12ZBeCj4JMqiu4eU7s97hQ@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Archived-At: <https://www.postgresql.org/message-id/089015d7-ba30-4113-a29f-248caca375f7%40aklaver.com>
Precedence: bulk

On 5/23/24 09:06, Skorpeo Skorpeo wrote:
> Thank you for the valuable feedback.  I see people are big fans of json 
> here.

You can be a fan of JSON and still think it is not the correct way to 
store data in a relational database. When you do that you often end up 
with nested, possibly unstructured, data sets. That ends up with the 
exertion of more effort to get the data out in any meaningful form and 
in a timely manner then the time and effort it would take to enter it in 
a structured way.

> 
> On Thu, May 23, 2024 at 3:04 AM Laurenz Albe <laurenz.albe@cybertec.at 
> <mailto:laurenz.albe@cybertec.at>> wrote:
> 
>     On Wed, 2024-05-22 at 22:38 -0500, Skorpeo Skorpeo wrote:
>      > I was wondering if having unrelated columns in a table is a sound
>     approach when
>      > using json.  In other words, if I have two collections of
>     unrelated json objects,
>      > for example "Users" and "Inventory", would it be ok to have one
>     table with a
>      > "Users" column and a "Inventory" column?  My concern is that from
>     a row
>      > perspective the columns could be different lengths, such as more
>     inventory
>      > items as users.  And for any given row the data in one column
>     would have no
>      > relation to another column. I would only query a single column at
>     a time.
>      >
>      > Would this approach be ok or are there pitfalls such that it would be
>      > advantageous/recommended to have a separate table for each column?
> 
>     It doesn't matter much if you use one or two columns.
>     But the word "collection" makes me worry.  Perhaps this article can give
>     you some ideas:
> 
>     https://www.cybertec-postgresql.com/en/json-postgresql-how-to-use-it-right/ <https://www.cybertec-postgresql.com/en/json-postgresql-how-to-use-it-right/>
> 
>     Yours,
>     Laurenz Albe
> 

-- 
Adrian Klaver
adrian.klaver@aklaver.com