agora inbox for pgsql-sql@postgresql.orghelp / color / mirror / Atom feed
foreign key referencing inheritance parent 3+ messages / 3 participants [nested] [flat]
* foreign key referencing inheritance parent @ 2020-07-20 14:44 J Lumby <johnlumby@hotmail.com> 0 siblings, 1 reply; 3+ messages in thread From: J Lumby @ 2020-07-20 14:44 UTC (permalink / raw) To: pgsql-sql@lists.postgresql.org This has come up before elsewhere e.g. https://stackoverflow.com/questions/26034752/postgresql-inheritance-and-foreign-key-referencing-pare... but I don't see any mention in any pg mailing list. postgresql permits creation of a foreign key referencing a table which is the parent of child tables via inheritance but enforces that every foreign key value must exist as pkey ONLY in the parent - it throws an ERROR otherwise, even if the pkey exists in a child table. This is problematic (to me) for three slightly different reasons : 1) it is surprising. The expectation is that a reference to accessing row(s) in a parent table will also search all children, as with SELECT, unless the ONLY keyword is specified. which then leads to ... 2) it is inconsistent with, and less useful than, the somewhat similar CHECK clause in a CREATE/ALTER TABLE : _____________________________________________________ CHECK ( expression ) [ NO INHERIT ] A constraint marked with NO INHERIT will not propagate to child tables. _____________________________________________________ For CHECK, the default is propagation to child tables unless explicitly prevented, which is consistent with SELECT and the ONLY qualifier. For FOREIGN KEY there is no optional qualifier to express "propagation" (i.e. propagation of the search for primary key) or "ONLY" and the behaviour is always "ONLY". 3) I am probably wrong but as far as I can tell this behaviour is not documented anywhere. Would there be any interest in providing a choice, e.g. an optional [ INHERIT ] on the REFERENCES clause, in a future release? Cheers, John ^ permalink raw reply [nested|flat] 3+ messages in thread
* Re: foreign key referencing inheritance parent @ 2020-07-31 07:07 Samed YILDIRIM <samed@reddoc.net> parent: J Lumby <johnlumby@hotmail.com> 0 siblings, 1 reply; 3+ messages in thread From: Samed YILDIRIM @ 2020-07-31 07:07 UTC (permalink / raw) To: J Lumby <johnlumby@hotmail.com>; pgsql-sql@lists.postgresql.org <pgsql-sql@lists.postgresql.org> <div><div>Hi John,</div><div> </div><div>Please find answers below.</div></div><div> </div><div>20.07.2020, 17:44, "J Lumby" <johnlumby@hotmail.com>:</div><blockquote><p>This has come up before elsewhere e.g.<br /><br /><a href="https://stackoverflow.com/questions/26034752/postgresql-inheritance-and-foreign-key-referencing-pare... /><br />but I don't see any mention in any pg mailing list.<br /><br /><br />postgresql permits creation of a foreign key referencing a table which<br />is the parent of child tables via inheritance<br /><br />but enforces that every foreign key value must exist as pkey ONLY in the<br />parent - it throws an ERROR otherwise, even if the pkey exists in a<br />child table.</p></blockquote><div>Having of primary key on child table with the same columns doesn't mean that it guarantees uniqueness of a record among parent and child tables. Each primary key provides uniqueness of records only in the table on which they are created.</div><blockquote><p><br />This is problematic (to me) for three slightly different reasons :<br /><br />1) it is surprising. The expectation is that a reference to<br />accessing row(s) in a parent table will also search all children, as<br />with SELECT, unless the ONLY keyword is specified.<br /><br />which then leads to ...</p></blockquote><div><br />The answer above is also valid here :)</div><blockquote><p><br /><br />2) it is inconsistent with, and less useful than, the somewhat<br />similar CHECK clause in a CREATE/ALTER TABLE :<br /><br />_____________________________________________________<br /><br /> CHECK ( expression ) [ NO INHERIT ]<br /><br />A constraint marked with NO INHERIT will not propagate to child tables.<br /><br />_____________________________________________________<br /><br />For CHECK, the default is propagation to child tables unless<br />explicitly prevented, which is consistent with SELECT and the ONLY<br />qualifier.<br /><br />For FOREIGN KEY there is no optional qualifier to express "propagation" <br />(i.e. propagation of the search for primary key) or "ONLY" and the<br />behaviour is always "ONLY".<br /><br /><br />3) I am probably wrong but as far as I can tell this behaviour is not<br />documented anywhere.</p></blockquote><div><a href="https://www.postgresql.org/docs/12/ddl-inherit.html#DDL-INHERIT-CAVEATS">https://www.postgres... href="https://www.postgresql.org/docs/12/ddl-partitioning.html">https://www.postgresql.org/docs/12/...; - Title 5.11.3.3. Caveats</div><blockquote><p><br /><br />Would there be any interest in providing a choice, e.g. an optional [<br />INHERIT ] on the REFERENCES clause, in a future release?</p></blockquote><div> </div><div>The most important part is why you need to use foreign key referencing to parent table of inheritance. Depending on your answer, there are multiple solution. For example, if the reason why you need it is that you use table partitioning implemented by using triggers and table inheritance, such as in the document below[1], you can switch declarative partitioning with PostgreSQL 12. PostgreSQL 12 supports foreign keys referencing to partitioned tables. [2]</div><div> </div><div>[1]: <a href="https://www.postgresql.org/docs/12/ddl-partitioning.html#DDL-PARTITIONING-IMPLEMENTATION-INHERITANCE...;[2]: <a href="https://www.postgresql.org/docs/12/sql-createtable.html">https://www.postgresql.org/docs/12/s... establishing a unique constraint for a multi-level partition hierarchy, all the columns in the partition key of the target partitioned table, as well as those of all its descendant partitioned tables, must be included in the constraint definition.</li><li><code>PRIMARY KEY</code> constraints share the restrictions that <code>UNIQUE</code> constraints have when placed on partitioned tables.</li></ul><div>If table partitioning is not the reason, the most generalized solution can be to use constraint triggers[3] rather than foreign keys. And you can guarantee uniqueness of records among parent and child tables by using check constraints and primary key/unique key constraints together.</div><div> </div><div>[3]: <a href="https://www.postgresql.org/docs/12/sql-createtrigger.html">https://www.postgresql.org/docs/12... />Cheers, John<br /><br /> </p></blockquote><div>Best regards.</div><div>Samed YILDIRIM</div> ^ permalink raw reply [nested|flat] 3+ messages in thread
* Re: foreign key referencing inheritance parent @ 2020-07-31 15:48 John Lumby <johnlumby@hotmail.com> parent: Samed YILDIRIM <samed@reddoc.net> 0 siblings, 0 replies; 3+ messages in thread From: John Lumby @ 2020-07-31 15:48 UTC (permalink / raw) To: Samed YILDIRIM <samed@reddoc.net>; pgsql-sql@lists.postgresql.org <pgsql-sql@lists.postgresql.org> On 07/31/20 03:07, Samed YILDIRIM wrote: > 20.07.2020, 17:44, "J Lumby" <johnlumby@hotmail.com>: > > but enforces that every foreign key value must exist as pkey ONLY > in the > parent - it throws an ERROR otherwise, even if the pkey exists in a > child table. > > Having of primary key on child table with the same columns doesn't > mean that it guarantees uniqueness of a record among parent and child > tables. Each primary key provides uniqueness of records only in the > table on which they are created. Of course! Thanks for pointing that out, I should have known it as pkeys are not inherited. > > > For CHECK, the default is propagation to child tables unless > explicitly prevented, which is consistent with SELECT and the ONLY > qualifier. > > For FOREIGN KEY there is no optional qualifier to express > "propagation" > (i.e. propagation of the search for primary key) or "ONLY" and the > behaviour is always "ONLY". > > > 3) I am probably wrong but as far as I can tell this behaviour > is not > documented anywhere. > > https://www.postgresql.org/docs/12/ddl-inherit.html#DDL-INHERIT-CAVEATS > https://www.postgresql.org/docs/12/ddl-partitioning.html - Title > 5.11.3.3. Caveats Ah I see. But it would be nice to reference those from CREATE TABLE and ALTER TABLE which is where someone is likely to be. > > > The most important part is why you need to use foreign key referencing > to parent table of inheritance. Depending on your answer, there are > multiple solution. For example, if the reason why you need it is that > you use table partitioning implemented by using triggers and table > inheritance, such as in the document below[1], you can switch > declarative partitioning with PostgreSQL 12. PostgreSQL 12 supports > foreign keys referencing to partitioned tables. [2] > [1]: > https://www.postgresql.org/docs/12/ddl-partitioning.html#DDL-PARTITIONING-IMPLEMENTATION-INHERITANCE > [2]: https://www.postgresql.org/docs/12/sql-createtable.html > > * When establishing a unique constraint for a multi-level partition > hierarchy, all the columns in the partition key of the target > partitioned table, as well as those of all its descendant > partitioned tables, must be included in the constraint definition. > * |PRIMARY KEY| constraints share the restrictions that |UNIQUE| > constraints have when placed on partitioned tables. > > Yes, we are using inheritance for table partitioning into time ranges, very similar to the examples in the reference manual. However, unfortunately, the primary key is not the partitioning column, and I *think* (correct me if wrong) that would disallow defining a referential constraint pointing to a declarative partition-set. In effect, what we need is for the referential constraint enforcement to scan all tables of the set in the same way as a SELECT. Having said that, I do see that declarative partitioning is much tidier and well-defined than inheritance+trigger so there are probably advantages in changing to that - but we would be no better off as regards referential constraints. > > > Cheers, John > > Best regards. > Samed YILDIRIM ^ permalink raw reply [nested|flat] 3+ messages in thread
end of thread, other threads:[~2020-07-31 15:48 UTC | newest] Thread overview: 3+ messages (download: mbox mbox.gz follow: Atom feed) -- links below jump to the message on this page -- 2020-07-20 14:44 foreign key referencing inheritance parent J Lumby <johnlumby@hotmail.com> 2020-07-31 07:07 ` Samed YILDIRIM <samed@reddoc.net> 2020-07-31 15:48 ` John Lumby <johnlumby@hotmail.com>
This inbox is served by agora; see mirroring instructions for how to clone and mirror all data and code used for this inbox