[postgis-users] PostGIS - server side optimization ( Split Data )

Jordi Ferran jordi.ferran at gmail.com
Wed Jan 12 07:37:34 PST 2005


Yes you're right, I misunderstand your comment. Using tables as
"cache" for retrieving the "simplified" version instead of dynamically
"simplify" it. You must take into consideration the cost:
1.- 650.000 geometries * 2 copies ( one simplified copy )
2.- SQL select/update/delete, manage both tables simulatenously
3.- the most important: you assume the simplify factor is fixed !

On my case the simplify factor is dynamic. I develop a Web Map
Service, so the bounding box and the output rendering image size is
delimited by the client.

Sample: a 8 kilometer box rendered on a 800 pixel image is a 10meter
per pixel, so it is impossible to draw a line that have a length lower
than 10 meters. As a consequence i can simplify the geometry with a
factor greater than 10 meter.

Anyway, thanks for you opinion (may be an option for other people).

Regards,
Jordi Ferran Sanchez

On Wed, 12 Jan 2005 18:45:18 +1300, Ralph Mason <ralph.mason at telogis.com> wrote:
> Jordi Ferran wrote:
> 
> >Hi,
> >
> >Thanks to all for your information. I must say that you helped me.
> >
> >- About data organization: split the data using multiple tables is not
> >a solution. If you need to use all that information does not matter
> >where the data is located.
> >
> >
> But if the other tables have a simplified copy of the data then it's not
> the same data.  Example if you are zoomed in you are probably interested
> in features down to a few thousandths of a degree.  At higher altitudes
> you don't really care and just need a simple outline of the poly (which
> is what is going to be drawn anyway).  You likely (depending on the size
> of your data) can't read and simplify all that data at the same speed
> that you could read pre simplified data.
> 
> Ralph
>



More information about the postgis-users mailing list