[postgis-users] Re: MySQL 4.1 GIS Functionality

[Paul DuBois]

At 10:42 -0700 8/5/03, Paul Ramsey wrote:
>On Tuesday, August 5, 2003, at 10:25 AM, Paul DuBois wrote:
>>
>>Did you compile MySQL yourself or use a precompiled distribution?
>>(Reason I ask is that we have observed some problems with GCC
>>and GIS columns unless optimization is turned off, or at least down
>>to -O1.)
>
>Ourselves, and it looks like the configure script chose to optimize 
>at O3. We did not have any problems with the columns themselves, 
>just thinks like the ISAM limitation (perhaps that was the cause of 
>the crash though?)

Possibly, although I wouldn't want to simply invoke "compiler error"
as the definitive cause and rule out possible bugs in the GIS implementation.

>>
>>You could also enable MySQL's query cache, in which case the second and
>>subsequent execution of a query should yield almost instantaneous
>>results. :-)
>
>Not actually calculating the answer isn't really fair. :) Nor is it 
>real-world for the spatial domain, where spatial queries can be 
>expected to almost *never* share the same input bounding box.

Agreed.

In MySQL's case, the effectiveness of the cache would vary depending
on how varied your queries are and how much memory you devote to the
cache.  If there are never any repeat queries, you'd actually see a
(slight) performance hit.

>
>      Paul Ramsey
>      Refractions Research
>      Email: pramsey at refractions.net
>      Phone: (250) 885-0632
>
>
>_______________________________________________
>postgis-users mailing list
>postgis-users at postgis.refractions.net
>http://postgis.refractions.net/mailman/listinfo/postgis-users