Optimize data for speed

[Brock Anderson]

Fabian,

Multinet data is divided by up state.  You'll want to use tile indexes 
for a lot of your layers.  That's in addition to having spatial indexes 
on the shapefiles.

There's a lot of processing that you *could* do on the data in the 
interest of making your maps render faster or in achieving certain 
cartographic effects.  The main principal you'll want to observe is this:

If a row in your shapefile is not being drawn (for a given map request), 
then it shouldn't be in the shapefile.  It's inefficient to make 
mapserver scan rows that it just ignores.

It's difficult to achieve that ideal for all map requests, but that's 
what you should aim for with any processing you do. 

Brock

Fabian Rami wrote:
> Hi all,
>
> I am working with "Tele Atlas multinet" data set.
>
> I try to figure out what processing must i do on this data to have 
> some rapid working data  for Mapserver.
>
> The data on the dvd contains a lot of file like that :
>
> belbel________00.dbf.gz
> belbel________00.prj.gz
> belbel________00.shp.gz
> belbel________00.shx.gz
>
> What must i do with all this file to have some optimized data for 
> mapserver.
>
> Aaron said to use "shptree" for having a spatially indexed shapefile.
> Is that all i have to do ?
>
> Can i make 1 big shapefile or it is better to have a lot of small 
> shapefile.
>
> I don't know how to do! So, i am waiting for all your precious comment 
> to help me to have a rapid and well configured mapserver :)
>
> Greetings,
>
> Fabian