Thread: Postgres Partitions Limitations (5.11.2.3)

The following documentation comment has been logged on the website:

Page: https://www.postgresql.org/docs/15/ddl-partitioning.html
Description:

Link:
https://www.postgresql.org/docs/current/ddl-partitioning.html#DDL-PARTITIONING-DECLARATIVE

"Using ONLY to add or drop a constraint on only the partitioned table is
supported as long as there are no partitions. Once partitions exist, using
ONLY will result in an error. Instead, constraints on the partitions
themselves can be added and (if they are not present in the parent table)
dropped." This seems in contradiction to the example involving adding a
unique constraint while minimizing locking at the bottom of "5.11.2.2.
Partition Maintenance", which seems to run fine on my local Pg instance:

"
This technique can be used with UNIQUE and PRIMARY KEY constraints too; the
indexes are created implicitly when the constraint is created. Example:

```ALTER TABLE ONLY measurement ADD UNIQUE (city_id, logdate);

ALTER TABLE measurement_y2006m02 ADD UNIQUE (city_id, logdate);
ALTER INDEX measurement_city_id_logdate_key
    ATTACH PARTITION measurement_y2006m02_city_id_logdate_key;
...
```
"

I might be misinterpreting something. Sorry if that's the case! 

Thanks,
Bryce

On Fri, 2023-01-06 at 08:28 +0000, PG Doc comments form wrote:
> The following documentation comment has been logged on the website:
>
> Page: https://www.postgresql.org/docs/15/ddl-partitioning.html
> Description:
>
> Link:
> https://www.postgresql.org/docs/current/ddl-partitioning.html#DDL-PARTITIONING-DECLARATIVE
>
> "Using ONLY to add or drop a constraint on only the partitioned table is
> supported as long as there are no partitions. Once partitions exist, using
> ONLY will result in an error. Instead, constraints on the partitions
> themselves can be added and (if they are not present in the parent table)
> dropped." This seems in contradiction to the example involving adding a
> unique constraint while minimizing locking at the bottom of "5.11.2.2.
> Partition Maintenance", which seems to run fine on my local Pg instance:
>
> "
> This technique can be used with UNIQUE and PRIMARY KEY constraints too; the
> indexes are created implicitly when the constraint is created. Example:
>
> ```ALTER TABLE ONLY measurement ADD UNIQUE (city_id, logdate);
>
> ALTER TABLE measurement_y2006m02 ADD UNIQUE (city_id, logdate);
> ALTER INDEX measurement_city_id_logdate_key
>     ATTACH PARTITION measurement_y2006m02_city_id_logdate_key;
> ...
> ```
> "
>
> I might be misinterpreting something. Sorry if that's the case!

No, that is actually an omission in the documentation.

The attached patch tries to improve that.

Yours,
Laurenz Albe