On 2019-Feb-04, Michael Paquier wrote:
> pg_partition_root() has not made it to the finish line yet, still it
> would have been nice to see a rebase, and the patch has been waiting
> for input for 4 weeks now. So I am marking it as returned with
> feedback.
Thanks for committing pg_partition_root ... but it turns out to be
useless for this purpose. It turns out that we need to obtain the list
of *ancestors* of the table being displayed, which pg_partition_tree
does not easily give you. So I ended up adding yet another auxiliary
function, pg_partition_ancestors, which in its current formulation
returns just a lits of OIDs of ancestor tables. This seems generally
useful, and can be used in conjunction with pg_partition_tree():
alvherre=# select t.* from pg_partition_tree(pg_partition_root('pk11')) t
join pg_partition_ancestors('pk11', true) a on (t.relid = a.relid);
relid | parentrelid | isleaf | level
-------+-------------+--------+-------
pk | | f | 0
pk1 | pk | f | 1
pk11 | pk1 | t | 2
(3 filas)
(A small tweak is to change the return type from OID to regclass.
Docbook additions missing also.)
Anyway, given this function, it's possible to fix the psql display to be
as I showed previously. Patches attached.
--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services