[mapserver-users] Tigerline

[Stephen Woodbridge]

Bruce, Ashley,

I haven't tried GDAL because my understanding was that using it to
access TIGER source data was pretty inefficent relative to shape files.
Especially if you plan to do more than one state. I use shape files. I
don't think you can use shptree since it only works with shape files.
tile4ms also reads a bunch of shape files and builds a tileindex. You
could in write your own utility to read the tiger source files and build
an equivalent tileindex but I don't think there is a utility already
available - you might want to check with Frank Koormann or scan his
posts in the list archive.

Bruce W Saunders/SDC/Systems/HSAS/MAXIMUS wrote:
> 
> I've been working with the tiger files and mapserver for a few months
> now (though I haven't made the indexes (but that certainly sounds like
> a good idea)). I've noticed an odd behavior for some fields with
> leading zeros (particularly zip codes and (I think) fips codes).  I
> think the problem is that the GDAL library treats these as integer and
> while parsing the file, drops leading zeros.  Thanks to the clarity of
> open source, I changed the fields to character, rebuilt GDAL and
> mapserver and things seemed to work fine.  The problem surfaced while
> I was trying to shade by zip code using an EXPRESSION in the mapfile.
> 
> -Bruce Saunders
> 
> Stephen Woodbridge <woodbri at swoodbridge.com>
> Sent by: owner-mapserver-users at lists.gis.umn.edu
> 02/27/2002 08:56 PM
> 
> To: Ashley Joyce <ashleyjoyce at hotmail.com>
> cc: "MAPSERVER (E-mail)" <mapserver-users at lists.gis.umn.edu>
> bcc:
> Subject: Re: [mapserver-users] Tigerline
> 
> Ashley,
> 
> Structure your data like this:
> 
> /mapdata/tiger2k/{SS}/{SS}{CCC}/{files}
> 
> where {SS} is the state fips code
> {CCC} is the fips county code
> {files} are all the files for that state+county
> 
> I would also put boundary files that are larger than a county and
> other
> datasets that cover the whole US in something like:
> 
> /mapdata/us
> /mapdata/canada
> /mapdata/mexico
> /mapdata/oceans
> etc
> 
> then create a tileindex for all the files under tiger2k and create
> shptree indexes for the tileindex and the files in the other
> directories. That should get you up and running!
> 
> You can see my version running with only the eastern half the US
> loaded
> at http://swoodbridge.com/web2/nmap it is using 9 GIGS
> 
> -Steve W.
> 
> > Ashley Joyce wrote:
> >
> > I have the mapserver demo version working properly and would like to
> > use data from the tigerline site for the USA. I will start with one
> > state at a time.... its heavy on space at around 3 -4 Gig. Can
> anyone
> > give me some advice on where to get some help or point to some
> > documentation for structuring the data files so it will work with
> > mapserver.... I'm still new to this product and haven't quite got my
> > head around all of it yet. I'm not very familiar with the GIS
> > structure. I'm running on a linux machine with apache.
> >
> > Thanks,
> >
> > Ashley Joyce