hans wulf wrote:
> Hi,
>
> if you want to do dirty counts or sums or any aggreate stuff, you will
> always have to visit the table. For many applications nobody cares
> about 0,01% inaccuracy.
>
> If you could keep the data that has to be aggregated in the index you
> could approximate values really fast.
>
> But because "Read uncommitted" is not implemented you will always have
> to visit the table. This is one reason why people have to still buy
> oracle.
>
> I don't know the postgres code, but I don't thing it is a big deal,
> not to care about consistancy. The code for executing such a query
> should be quite basic, because no MVCC-Stuff has to be done.
>
> Will this feature come any time soon? Even if "Read uncommitted" is a
> "could read all sorts of old and dirty stuff" it is still better than
> nothing.
Dirty reads are unlikely to be implemented. We do have a TODO item and
wiki page about how to allow index scans without heap access:
http://wiki.postgresql.org/wiki/Index-only_scans
-- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB
http://enterprisedb.com
+ It's impossible for everything to be true. +