Re: proposal: be smarter about i/o patterns in index scan

On Wed, 2004-05-19 at 07:58, Tom Lane wrote:
> "Jeffrey W. Baker" <jwbaker(at)acm(dot)org> writes:
> > We have noticed a way to make a major improvement in Pg's performance on
> > our workload, and I would like to get your thoughts before I go off to
> > work up a patch.
>
> For starters, read the previous discussions of this in the archives.

If you are referring to archives.postgresql.org, all I have to say is:

"An error occured! Can not connect to search daemon"

And as far as I have seen, it's been like that for years.

> Two questions you should have answers to before starting to implement,
> rather than after, are how to deal with locking considerations and
> what will be the implications of giving up the property that indexscans
> deliver sorted output.

Are you saying that index scan results are sorted by something other
than the index key? Because in my scheme they would still be sorted by
index key.

-jwb