Re: Incremental View Maintenance, take 2

On Mon, 28 Aug 2023 02:49:08 +0900
Yugo NAGATA <nagata(at)sraoss(dot)co(dot)jp> wrote:

> On Sun, 2 Jul 2023 10:38:20 +0800
> jian he <jian(dot)universality(at)gmail(dot)com> wrote:

I attahed the patches v29 updated to comments from jian he.
The changes from the previous includes:

- errors in documentations is fixed.
- remove unnecessary BEGIN from the test
- add isimmv column to pg_matviews system view
- fix a typo
- rebase to the master branch

>
> > ok. Now I really found a small bug.
> >
> > this works as intended:
> > BEGIN;
> > CREATE INCREMENTAL MATERIALIZED VIEW test_ivm AS SELECT i, MIN(j) as
> > min_j FROM mv_base_a group by 1;
> > INSERT INTO mv_base_a select 1,-2 where false;
> > rollback;
> >
> > however the following one:
> > BEGIN;
> > CREATE INCREMENTAL MATERIALIZED VIEW test_ivm1 AS SELECT MIN(j) as
> > min_j FROM mv_base_a;
> > INSERT INTO mv_base_a select 1, -2 where false;
> > rollback;
> >
> > will evaluate
> > tuplestore_tuple_count(new_tuplestores) to 1, it will walk through
> > IVM_immediate_maintenance function to apply_delta.
> > but should it be zero?
>
> This is not a bug because an aggregate without GROUP BY always
> results one row whose value is NULL.
>
> The contents of test_imv1 would be always same as " SELECT MIN(j) as min_j
> FROM mv_base_a;", isn't it?
>
>
> Regards,
> Yugo Nagata
>
> --
> Yugo NAGATA <nagata(at)sraoss(dot)co(dot)jp>
>
>

--
Yugo NAGATA <nagata(at)sraoss(dot)co(dot)jp>