Horton

Re: index-only scans

Hans-Jürgen Schönig <postgres@cybertec.at>

View thread
From: PostgreSQL - Hans-Jürgen Schönig <postgres@cybertec.at>
To: Heikki Linnakangas <heikki.linnakangas@enterprisedb.com>
Cc: Robert Haas <robertmhaas@gmail.com>, pgsql-hackers@postgresql.org
Date: 2011-08-12T20:21:25Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Reduce the alignment requirement of type "name" from int to char, and arrange

On Aug 12, 2011, at 10:03 PM, Heikki Linnakangas wrote:

> On 11.08.2011 23:06, Robert Haas wrote:
>> Comments, testing, review appreciated...
> 
> I would've expected this to use an index-only scan:
> 
> postgres=# CREATE TABLE foo AS SELECT generate_series(1,100000) AS id;
> SELECT 100000
> postgres=# CREATE INDEX i_foo ON foo (id) WHERE id = 10;
> CREATE INDEX
> postgres=# VACUUM ANALYZE foo;
> VACUUM
> postgres=# EXPLAIN SELECT id FROM foo WHERE id = 10;
>                           QUERY PLAN
> -----------------------------------------------------------------
> Index Scan using i_foo on foo  (cost=0.00..8.27 rows=1 width=4)
>   Index Cond: (id = 10)
> (2 rows)
> 
> If it's not a predicate index, then it works:
> 
> postgres=# DROP INDEX i_foo;
> DROP INDEX
> postgres=# EXPLAIN SELECT id FROM foo WHERE id = 10;
>                              QUERY PLAN
> -----------------------------------------------------------------------
> Index Only Scan using i_foo2 on foo  (cost=0.00..8.28 rows=1 width=4)
>   Index Cond: (id = 10)
> (2 rows)


is there any plan to revise the cost for index only scans compared to what it is now?

	many thanks,

		hans

--
Cybertec Schönig & Schönig GmbH
Gröhrmühlgasse 26
A-2700 Wiener Neustadt, Austria
Web: http://www.postgresql-support.de