Message-ID: <14719fdf-108a-edb6-fd74-0c42669bcfb4@gmail.com>
Date: Mon, 2 Oct 2023 21:14:16 -0600
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.15.1
Subject: Re: A situation with one parent table and 3 child tables
Content-Language: en-CA
To: pgsql-sql@lists.postgresql.org
References: 
 <CAAY=A7_wHBKkKFxuxXc7A+fr3TYGmWS9-hCGghLRDcQAAa_wfQ@mail.gmail.com>
 <CAKFQuwZbmqeLoHBzuOkVjd4qXyNUi1tcuhf8bECK+2c-DGK9iw@mail.gmail.com>
 <2206102.1696302184@sss.pgh.pa.us>
From: Rob Sargent <robjsargent@gmail.com>
In-Reply-To: <2206102.1696302184@sss.pgh.pa.us>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Archived-At: 
 <https://www.postgresql.org/message-id/14719fdf-108a-edb6-fd74-0c42669bcfb4%40gmail.com>
Precedence: bulk

On 10/2/23 21:03, Tom Lane wrote:
> "David G. Johnston" <david.g.johnston@gmail.com> writes:
>> On Monday, October 2, 2023, JORGE MALDONADO <jorgemal1960@gmail.com> wrote:
>>> I have one parent table (*table_p*) with 3 child tables (*table_ch1*, *table_ch2
>>> *and *table_ch3*). Each record of the parent table can be associated with
>>> 1 and only 1 child table records. This means that:
>>>
>>> * Some records of the *table_p* will link to records of *table_ch1*
>>> * Some records of the *table_p* will link to records of *table_ch2*
>>> * Some records of the *table_p* will link to records of *table_ch3*
>>>
>>> At first look, this does not make very much sense to me. I thought about
>>> considering 3 parent tables, one for each child table. However, the 3
>>> parent tables would have the same exact structure and I would like to know
>>> if there is a workaround for this issue.
>> You are thinking of it backwards.  Your chN tables will have FK pointing
>> back to the p table.  I suggest adding some kind of type column to the p
>> table indicating which chN table the row belongs to.
> Do you need that?  I was wondering about converting the 3 child tables
> into a partitioned table.  Then you can query them separately when
> you need to, but you can also treat them as one table --- and you
> can set up one FK constraint between that and the parent table.
>
> 			regards, tom lane
>
>

I've implemented David's suggestion in the past.  Presumably the three 
child tables are of different structure but share the attributes of the 
"parent"