On Tue, 13 Oct 2009, dan wrote: > Let's say I have a table t with 5 columns c1 NOT NULL, c2 NOT NULL, c3, c4, > c5 > and I have a UNIQUE index on (c1, c2) (remember c1 and c2 have a not null > constraint) > > When I run the query: > select c1,c2 from t > > I expect the explain to say index scan; instead it says table scan. > The index has ALL the info I need and in my case is about 2% the size of the > table.
The index is missing the visibility information, so right now the system cannot run such a query over the index alone since it couldn't determine if the row version is visible to your statement. That data is currently in with the data in the table, so it ends up needing to go back and read the rows from the main table. IIRC, there's been talk and some work around supporting indexes with visibility info, but I think there might have been some issues that needed to be worked out before that would work. -- Sent via pgsql-bugs mailing list (pgsql-bugs@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-bugs
