Re: Defer a functional index calculation?
| От | Randall Lucas |
|---|---|
| Тема | Re: Defer a functional index calculation? |
| Дата | |
| Msg-id | 7d5145af0903241254s154da2d6r4eaec2eb14dded93@mail.gmail.com обсуждение |
| Ответ на | Defer a functional index calculation? (Randall Lucas <rlucas@tercent.com>) |
| Список | pgsql-general |
On Mon, Mar 23, 2009 at 4:46 AM, Sam Mason <sam@samason.me.uk> wrote:
Busted! Yes, I was lying to postgres. (The function is all-but-immutable, your honor; I was only trying lazily to memoize its output...)
Looks like that's what I'm headed for.
Thank you,
Randall
On Fri, Mar 20, 2009 at 05:37:33PM -0700, Randall Lucas wrote:I don't think you can do that; are you lying in the function's
> I added a functional index.
>
> create table example (id serial primary key, stuff text, parent_id int);
> create index example_root_idx on example (get_root_id(id));
>
> (get_root_id(id) pulls an example row and recurses onto parent_id until it
> hits a root)
definition that it's "immutable"? As far as I know, and a quick check
Busted! Yes, I was lying to postgres. (The function is all-but-immutable, your honor; I was only trying lazily to memoize its output...)
What about having some trigger to "cache" the entry's root "parent_id"
in another column?
Looks like that's what I'm headed for.
Thank you,
Randall
В списке pgsql-general по дате отправления: