1

I've downloaded the geocommons/geocoder source and have one small sample TigerLine zip file from the census site saved into /opt/tiger/tl_2010_01_state10.zip

I've tried to run the tiger_import tool on this file with the command:

build/tiger_import /opt/tiger/geocoder.db /opt/tiger

with all of the prerequisite gems installed, specifically: Text, fastercsv and sqlite3-ruby gems as well as running make and make install.

However, when I execute tiger_import, I get the error:

ls: /opt/tiger/*/*/tl_*_edges.zip: No such file or directory

although there seems to be a geocoder.db file created in /opt/tiger.

Does any have better information on the steps necessary to build the tiger lines data with the geocoder?

j0k
  • 22,600
  • 28
  • 79
  • 90
Jack Flemming
  • 11
  • 2

1 Answers1

1

The script expects a directory structure more like the 2009 data: ftp://ftp2.census.gov/geo/tiger/TIGER2009/

Under your tiger directory, you'll need one state (01_ALABAMA, for instance) and one county (01001_Autauga_County), and inside that you'll need the _addr.zip, _edges.zip, and _featnames.zip files.

It's true: the 2010 data isn't set up this way (there's a giant directory for each shape type, and a file for each county in those directories) but the import script isn't set up to use the 2010 data as it's written now. Given the way the script is set up, you might have less heartache using the 2009 data until the import scripts get updated for the 2010 data. Or until the 2010 all gets published.

Boeckm
  • 3,264
  • 4
  • 36
  • 41
Bill
  • 11
  • 1