Re: B-tree parent pointer and checkpoints

On Fri, Nov 12, 2010 at 7:20 PM, Heikki Linnakangas
<heikki.linnakangas@enterprisedb.com> wrote:
> I think we can work around that with a small modification to the page split
> algorithm. In a nutshell, when the child page is split, put a flag on the
> left half indicating that the rightlink must always be followed, regardless
> of the NSN. When the downlink is inserted to the parent, clear the flag.
> Setting and clearing of these flags need to be performed during WAL replay
> as well.
>

Does this not cause duplicate results? Or does GIST already have to be
prepared to deal with duplicate results?


-- 
greg