Re: delta relations in AFTER triggers

Kevin Grittner <kgrittn(at)ymail(dot)com> wrote:

> I think this is approaching a committable state, although I think
> it should probably be broken down to four separate patches.

And here they are. This should net to the same set of changes as
the prior post, but the changes are logically separated. They are
labeled as v3 to match the last post.

trigger-transition-tables allows transition table names to be
specified in a REFERENCING clause of CREATE TRIGGER, per spec, and
creates tuplestores when needed in the TriggerData structure. It
doesn't worry about who does what with that data. This doesn't
depend on anything else.
15 files changed, 577 insertions(+), 43 deletions(-)

spi-tuplestore-registry allows tuplestores, with associated name
and TupleDesc, to be registered with the current SPI connection.
Queries planned or executed on that connection will recognize the
name as a tuplestore relation. It doesn't care who is registering
the tuplestores or what happens to them. It doesn't depend on
anything else.
5 files changed, 445 insertions(+)

executor-tuplestore-relations covers parse analysis,
planner/optimizer, and executor layers. It pulls information from
the registry in a couple places, but is not very intertwined with
SPI. That is the only registry it knows right now, but it should
be easy to add other registries if needed. It doesn't care where
the tuplestore came from, so we should be able to use this for
other things. I have it in mind to use it for incremental
maintenance of materialized views, but I expect that other uses
will be found. It has a logical dependency on the
spi-tuplestore-registry patch. While it doesn't have a logical
dependency on trigger-transition-tables, they both modified some of
the same files, and this won't apply cleanly unless
trigger-transition-tables is applied first. If you hand-correct
the failed hunks, it compiles and runs fine without
trigger-transition-tables.
30 files changed, 786 insertions(+), 9 deletions(-)

plpgsql-after-trigger-transition-tables takes the tuplestores from
TriggerData and registers them with SPI before trigger planning and
execution. It depends on the trigger-transition-tables and
spi-tuplestore-registry patches to build, and won't do anything
useful at run time without the executor-tuplestore-relations patch.
3 files changed, 37 insertions(+), 11 deletions(-)

Hopefully this will make review easier.

--
Kevin Grittner
EDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company