On Tue, May 7, 2024 at 7:54 AM Bruce Momjian <br...@momjian.us> wrote:
> In this commit:
>
> commit 34768ee3616
> Author: Peter Eisentraut <pe...@eisentraut.org>
> Date: Sun Mar 24 07:37:13 2024 +0100
>
> Add temporal FOREIGN KEY contraints
>
> Add PERIOD clause to foreign key constraint definitions. This
> is
> supported for range and multirange types. Temporal foreign
> keys check
> for range containment instead of equality.
>
> This feature matches the behavior of the SQL standard temporal
> foreign
> keys, but it works on PostgreSQL's native ranges instead of
> SQL's
> "periods", which don't exist in PostgreSQL (yet).
>
> Reference actions ON {UPDATE,DELETE} {CASCADE,SET NULL,SET
> DEFAULT}
> are not supported yet.
>
> Author: Paul A. Jungwirth <p...@illuminatedcomputing.com>
> Reviewed-by: Peter Eisentraut <pe...@eisentraut.org>
> Reviewed-by: jian he <jian.universal...@gmail.com>
> Discussion:
> https://www.postgresql.org/message-id/flat/CA+renyUApHgSZF9-nd-a0+OPGharLQLO=mdhcy4_qq0+noc...@mail.gmail.com
>
> this text was added to create_table.sgml:
>
> In addition, the referenced table must have a primary
> key or unique constraint declared with <literal>WITHOUT
> --> OVERLAPS</literal>. Finally, if one side of the foreign key
> --> uses <literal>PERIOD</literal>, the other side must too. If the
> <replaceable class="parameter">refcolumn</replaceable> list is
> omitted, the <literal>WITHOUT OVERLAPS</literal> part of the
> primary key is treated as if marked with <literal>PERIOD</literal>.
>
> In the two marked lines, it says "if one side of the foreign key uses
> PERIOD, the other side must too." However, looking at the example
> queries, it seems like if the foreign side has PERIOD, the primary side
> must have WITHOUT OVERLAPS, not PERIOD.
>
> Does this doc text need correcting?
>
>
The text is factually correct, though a bit hard to parse.
"the other side" refers to the part after "REFERENCES":
FOREIGN KEY ( column_name [, ... ] [, PERIOD column_name ] ) REFERENCES
reftable [ ( refcolumn [, ... ] [, PERIOD column_name ] ) ]
***(shouldn't the second occurrence be [, PERIOD refcolum] ?)
The text is pointing out that since the refcolumn specification is optional
you may very well not see a second PERIOD keyword in the clause. Instead
it will be inferred from the PK.
Maybe:
Finally, if the foreign key has a PERIOD column_name specification the
corresponding refcolumn, if present, must also be marked PERIOD. If the
refcolumn clause is omitted, and thus the reftable's primary key constraint
chosen, the primary key must have its final column marked WITHOUT OVERLAPS.
David J.
