Re: show Heap Fetches in EXPLAIN for index-only scans
Robert Haas <robertmhaas@gmail.com>
From: Robert Haas <robertmhaas@gmail.com>
To: Peter Geoghegan <peter@2ndquadrant.com>
Cc: Magnus Hagander <magnus@hagander.net>, pgsql-hackers@postgresql.org
Date: 2012-01-13T16:08:16Z
Lists: pgsql-hackers
On Fri, Jan 13, 2012 at 10:41 AM, Peter Geoghegan <peter@2ndquadrant.com> wrote: > On 13 January 2012 15:31, Robert Haas <robertmhaas@gmail.com> wrote: >> On Fri, Jan 13, 2012 at 10:29 AM, Magnus Hagander <magnus@hagander.net> wrote: >>> Would also be good to have counter sin pg_stat_* for this, since you'd >>> usually want to look at this kind of data over time as well. In your >>> plans? ;) >> >> Not really. I don't have a clear enough idea about what that should >> look like, and I expect a vigorous debate over the distributed cost of >> another counter. But I'm happy to have someone who feels more >> strongly about it than I do take up the cause. > > Maybe the the ioss_HeapFetches field should be a uint32? That's all > it's going to be on some platforms anyway. I originally had it that way, but then it didn't match what ExplainPropertyLong() was looking for. I didn't think it was worth further complicating explain.c for this... -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company