Re: Index-only scans

From: Ron Mayer <rm_pg(at)cheapcomplexdevices(dot)com>
To: Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com>
Cc: Simon Riggs <simon(at)2ndQuadrant(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Index-only scans
Date: 2009-07-15 00:21:26
Message-ID: 4A5D2106.4040508@cheapcomplexdevices.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Heikki Linnakangas wrote:
> ...
> CREATE TABLE manytomany (aid integer, bid integer);
> CREATE INDEX a_b ON manytomany (aid, bid);
> CREATE INDEX b_a ON manytomany (bid, aid);
> ...
>> new and interesting indexing strategies. Covered indexes are also one
>> kind of materialized view. It may be better to implement mat views and
>> gain wider benefits too.
>
> Materialized view sure would be nice, but doesn't address quite the same
> use cases. Doesn't help with the many-to-many example above, for
> example. We should have both.

Really? I'd have thought that index is similar to materializing
these views:
create view a_b as select aid,bid from manytomany order by aid,bid;
create view b_a as select bid,aid from manytomany order by bid,aid;
Or perhaps
create view a_b as select aid,array_agg(bid) from manytomany group by aid;

But I like the index-only scan better anyway because I already have
the indexes so the benefit would come to me automatically rather than
having to pick and choose what views to materialize.

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Jeremy Kerr 2009-07-15 00:42:07 Re: [PATCH 1/2 v3] [libpq] rework sigpipe-handling macros
Previous Message Robert Haas 2009-07-15 00:18:03 CommitFest 2009-07 is Now Closed