Re: Planner create a slow plan without an available index

Tom Lane wrote:
> Martijn van Oosterhout <kleptog@svana.org> writes:
>
>>rtree works on multidimesional (geometric) data. It can do range tests
>>(is object A to the left of object B) but it's only applicable if your
>>conditions can be interpreted that way.
>
>
>>GiST is for creating custom index types, hardly likely to be useful
>>in your case.
>
>
> Actually either rtree or GIST should be able to do something useful with
> this, since it's basically a 1-D overlap query.  The main problem with
> GIST is to find a suitable opclass, since there aren't any in the core
> system.  Possibly contrib/seg could be used.
>
>             regards, tom lane

Ok first of all thanks guys as always for your help, and I will try to
use rtree to improve my query (hopefuly ill be able to come back and say
that it worked :)).

I got another question which is not connected and probably its just me
being paranoid late at night but still... :)

at chapter "21.1.3. Preventing transaction ID wraparound failures" of
the postgresql manual its written the following info:
"Prior to PostgreSQL 7.2, the only defense against XID wraparound was to
re-initdb at least every 4 billion transactions. This of course was not
very satisfactory for high-traffic sites, so a better solution has been
devised. The new approach allows a server to remain up indefinitely,
without initdb or any sort of restart. The price is this maintenance
requirement: every table in the database must be vacuumed at least once
every billion transactions."
My postgresql version is 8.01 (I should have mentioned that at start no? :))

Now again im probably just paranoid but when I'm starting a transaction
and in it im making more then 4 billions diffrent queries
(select,insert,update,truncate...) and then im closing it, its counted
as only one transaction right? (should I duck to avoid the manual? ;))

As always thanks alot!
Ben-Nes Yonatan