Single Index Tuple Chain (SITC) method

bruce wrote:
> Greg Stark wrote:
> >
> > Bruce Momjian <bruce(at)momjian(dot)us> writes:
> >
> > > PFC wrote:
> > > >
> > > > > My idea is that if an UPDATE places the new tuple on the same page as
> > > > > the old tuple, it will not create new index entries for any indexes
> > > > > where the key doesn't change.
> > > >
> > > > Basically the idea behind preventing index bloat by updates is to have
> > > > one index tuple point to several actual tuples having the same value.
> > > >
> > >
> > > The idea is not to avoid index bloat, but to allow heap reuse, and having
> > > one index entry for multiple versions of an UPDATEd row is merely an
> > > implementation detail.
> >
> > It sort of sounds like you're describing a whole new index type that stores
> > only the page, not the precise record of any tuple it indexes. If your table
>
> Background, indexes point to page item pointers, not to actual offsets
> in the page. This is how vacuum can move around tuples without modifying the
> indexes. The index points to a page item pointer that is a chain of
> tuples with the same indexed columns.

Here is an overview of the SITC method:

http://momjian.us/cgi-bin/pgsitc

Anyone want to start coding?

--
Bruce Momjian bruce(at)momjian(dot)us
EnterpriseDB http://www.enterprisedb.com

+ If your life is a hard drive, Christ can be your backup. +