Igor Shevchenko <[EMAIL PROTECTED]> writes:
> In both cases, tables are filled with ~10m of rows, "is_read" is false in the
> 1st case, and "NULL" in the 2nd. I did "VACUUM FULL ANALYSE" after both
> imports.
> Here's the problem: in the 2nd case, planner wouldn't choose an index scan
> using pa
On Mon, Dec 13, 2004 at 20:18:57 +0200,
Igor Shevchenko <[EMAIL PROTECTED]> wrote:
>
> Here's the problem: in the 2nd case, planner wouldn't choose an index scan
> using partial index on "is_read" for the following queries:
>
> explain select * from user_msg where is_read=true;
> explain selec
Hi all,
PostgreSQL v8.0.0rc1, two variants of a "user_msg" table:
create table user_msg (
message_id integer not null references message(id) on update cascade on
delete cascade,
user_id integer not null,
status smallint not null default 0,
is_read boolean not null default false,
unique
