[postgis-users] tiger data pain

[Stephen Woodbridge]

On 12/13/2011 10:16 PM, Andy Colson wrote:
> I downloaded PostGis 2 to play around with, and tried out the tiger
> stuff. I have to say, that's kinda painful.
>
> The doc says: add "set -e -u", which causes as many problems as it might
> catch.
> Editing the .sql was a little confusing, and the resulting .sh file is
> not really re-startable.
>
> Would anyone be interested in a perl script to simplify things? I'm
> thinking of replacing the stored proc that generates a script, all of
> which you have to edit, with a perl script (which would have a config to
> edit).
>
> Or... maybe other's have not had the problems I had? I don't really have
> a use for this right now, just playing around. If anyone had some
> requests, I might be interested in working on it.

Andy,

I think this would be a good idea. One trick I use a lot with Perl is 
find2perl which I use to walk a directory structure like the Tiger 
source tree to find the files I want to load.

Also Census just released the Tiger2011 files, unfortunately, they have 
only released them in the type directory structure and not in the state 
directory structure. Each year there are file name changes and column 
name changes within the files. I think using Perl would make it easier 
to remap these column names.

-Steve