On 2018-Jun-07, Ashutosh Bapat wrote:
> On Thu, Jun 7, 2018 at 10:58 AM, Amit Langote
> <Langote_Amit_f8@lab.ntt.co.jp> wrote:
> > I don't understand why you think it's too troublesome to let the users
> > know that there is some way to use BR triggers with partitioning. We
> > didn't do that for indexes, for example, before PG 11 introduced the
> > ability to create them on partitioned tables.
>
> By looking at the index keys it's easy to decide whether the two
> indexes are same. When we add an index on a partitioned table in v11,
> we skip creating an index on the partition if there exists an index
> similar to the one being created. So, a user can have indexes on
> partitions created in v10, upgrade to v11 and create an index on the
> partitioned table. Nothing changes. But that's not true for a trigger.
> It's not easy to check whether two triggers are same or not unless the
> underlying function is same. User may or may not be using same trigger
> function for all the partitions, which is more true, if the column
> order differs between partitions. So, if the user has created triggers
> on the partitions in v10, upgrades to v11, s/he may have to drop them
> all and recreate the trigger on the partitioned table.
Actually, the column order doesn't matter for a trigger function,
because these don't refer to columns by number but by name. So unless
users write trigger functions in C and use hardcoded column numbers
(extremely unlikely), I think this is not an issue. In other words, in
the interesting cases the trigger functions are the same for all
partitions -- therefore upgrading from separate per-partition triggers
to one master trigger in the partitioned table is not going to be that
difficult, ISTM.
--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services